ELF>@( @8@T T    & & 8 0 0 &0 &$$Ptdppp44QtdRtd  & &GNUY"I/[JGS5՛|w% B `"p @@ 0 ,@@D:`@nIHdCC1 (0b8Kk}LJ f|8\5M3ndy'IK}U=v^a#2ח*X3I#K霳N yME>=g!ni焽dMD.ۗqX|EڋFy$vWJ@IoS:rKffBE,_̩*1޺KFmK4bs\rb(zK|M&E @ 4 @ M,\   p _     `  _ |4 x N   #  Rhb /Lyy L  I rv u57 O1XmX /B c3 a >" & 1K K8 :R"%L  b B = PX Ж>  0Wh  \  P9 W7 ЏA Ю- ~  0 ` ` p/G $  \W    < @(n  @f _  '  ) 6 P p=   (  m  v '` $ 0)0 H) Fn [  @&   еPx  `mp , u p/9   0) @C `): @  -;  pr ~  & q p V @  PR   }__gmon_start___init_fini_ITM_deregisterTMCloneTable_ITM_registerTMCloneTable__cxa_finalize_Jv_RegisterClasses_Py_NoneStruct_PyThreadState_CurrentPyExc_TypeErrorPyErr_FormatPyExc_ValueErrorPyNumber_IntPyNumber_LongPyErr_OccurredPyErr_SetStringPyLong_AsUnsignedLongLongPyExc_OverflowError_Py_TrueStruct_Py_ZeroStructPyObject_IsTruePyErr_NormalizeException__stack_chk_failPyNumber_IndexPyInt_AsSsize_tPyExc_StopIterationPyErr_GivenExceptionMatchesrk_intervalmemcpyPyObject_GC_UnTrackPyErr_FetchPyMem_FreePyErr_WriteUnraisablePyErr_RestorePyObject_GetAttrPyModule_GetDictPyFrame_NewPyTraceBack_HerePyString_FromStringPyCode_NewPyMem_ReallocPyString_FromFormatPyMem_MallocPyLong_AsUnsignedLongPyLong_AsLongPyInt_FromSsize_tPyObject_SetItemPyObject_GetItemPyExc_SystemErrorPyType_IsSubtypePyImport_ImportPyOS_snprintfPyErr_WarnExPyDict_NextPyString_TypePyString_AsString_PyString_EqPyUnicodeUCS4_ComparePyTraceBack_TypePyExc_BaseExceptionPyExc_AttributeErrorPyErr_ExceptionMatchesPyExc_ImportErrorPyExc_NameErrorPyDict_GetItemPyObject_CallPyTuple_NewPyDict_NewPyInt_FromLongPyObject_CallFunctionObjArgsPyList_NewPyDict_SizePyDict_SetItemPyCObject_FromVoidPtrPyTuple_TypePyNumber_AddPyString_FromStringAndSizePyEval_SaveThreadrk_fillPyEval_RestoreThreadPyInt_Typerk_longPyLong_TypePyObject_IsInstancePyFloat_FromDoublerk_standard_cauchyrk_standard_exponentialrk_gaussrk_doublePyList_TypePyObject_RichComparePyObject_SizePyFloat_TypePyFloat_AsDoublePyObject_GetIterPyCObject_AsVoidPtrPyNumber_Subtractrk_random_uint8rk_random_uint64PyLong_FromUnsignedLongrk_random_uint16rk_random_uint32rk_random_boolPyErr_Clearrk_rayleighrk_powerrk_weibullrk_paretork_standard_trk_chisquarerk_standard_gammark_exponentialrk_binomialrk_logseriesrk_geometricrk_zipfrk_poissonPyNumber_Orinit_by_arrayrk_seedrk_randomseedPyNumber_MultiplyPyNumber_AndPyNumber_InPlaceAddPyList_AsTuplePyList_AppendPyObject_SetAttrPySequence_ContainsPyNumber_Remainderrk_hypergeometricPyInt_AsLongPyGILState_EnsurePyExc_ZeroDivisionErrorPyGILState_Releaserk_triangularrk_noncentral_frk_waldrk_lognormalrk_logisticrk_gumbelrk_laplacerk_vonmisesrk_noncentral_chisquarerk_frk_gammark_betark_normal__finiterk_uniform_Py_EllipsisObjectrk_negative_binomialPyNumber_InPlaceDividePySlice_NewinitmtrandPy_GetVersionPy_InitModule4_64PyImport_AddModulePyObject_SetAttrStringPyString_InternFromStringPyUnicodeUCS4_DecodeUTF8PyInt_FromString__pyx_module_is_main_mtrandPyTuple_PackPyType_ReadyPyCapsule_NewPyDict_SetItemStringPyImport_ImportModulePyObject_GetAttrStringPyCObject_TypePyExc_RuntimeErrorPyErr_PrintPyCFunction_NewExPyType_Modifiedrk_randomrk_ulongrk_devfillfopen64freadfclosegettimeofdaygetpidclockrk_altfilllogsqrtrk_strerrorpowexprk_binomial_btpefloorrk_binomial_inversionrk_poisson_multrk_poisson_ptrsacosfmodrk_geometric_searchrk_geometric_inversionceilrk_hypergeometric_hyprk_hypergeometric_hrualibpython2.7.so.1.0libpthread.so.0libc.so.6_edata__bss_start_end/opt/alt/python27/lib64GLIBC_2.14GLIBC_2.4GLIBC_2.2.5 M ii X ui b  &P &( &( &n))n))n))n)x)n)X)n))n))o))o))o)()o))@o)x)Ho)X)Po))`o)x)ho)X)po))o))o))o))o))o))o))o))o))o))p))p))p))@p))Hp))Pp))`p))hp))pp))p))p))p))p))p))p))p))p))p))q))q))q))@q)P)Hq))Pq))`q))hq))q))q))q))q))q))q))q))q))q))q))r))r))r)) r)h)(r))0r))@r))Hr))`r)h)hr))pr))r))r)`)r))r))r)8)r))r))r)8)r))r))s))s)) s))(s))@s))Hs))`s))hs)()ps))xs))s))s))s)h)s)8)s))s)8)s))s))t))t)) t))0t))8t) )`t)x)ht)X)pt))xt))t)x)t)X)t))t))t)x)t)X)t))t)) u)x)(u)X)0u))8u))`u)x)hu)X)pu))xu))u)x)u)X)u))u))u)x)u)X)u))u)) v)x)(v)X)0v))8v))`v)x)hv)X)pv))xv))v))v))')r)\)' )N()G8)'@)H)X) '`)h)xx)'))@o) '))`)')(ȵ)5ص) '))e)')) )L()U &(&0&8&@&H&P&X& `&h& p& x& &&&&&&&&&+&&3&&&4&&=&>&&& &(&J0&K8&N@&H&P&TX&Z`&[h&\p&x&`&&a&j&k&&s&t&v&&w&~&&&&&& &(&0&8&@&H&P& X&`&h&p&x&&&&&&&&&&&&&&&&&&&&& & (&!0&"8&@&#H&P&$X&%`&&h&p&'x&(&)&*&,&&-&.&/&0&&1&2&&5&6&7&&&8&9&: &;(&<0&8&@&H&?P&@X&A`&Bh&p&Cx&D&E&&&F&G&H&&I&K&L&M&O&&P&Q&&R&S&U&V &W(&X0&8&@&H&YP&]X&^`&_h&p&x&b&c&d&e&f&g&h&i&l&&m&n&o&p&q&r&u&x&&y&z &{(&0&|8&}@&H&P&X&`&h&p&x&&&&&&&HHu%HtkH5v%%v%@%v%h%v%h%v%h%v%h%v%h%zv%h%rv%h%jv%hp%bv%h`%Zv%h P%Rv%h @%Jv%h 0%Bv%h %:v%h %2v%h%*v%h%"v%h%v%h%v%h% v%h%v%h%u%h%u%h%u%hp%u%h`%u%hP%u%h@%u%h0%u%h %u%h%u%h%u%h%u%h %u%h!%u%h"%u%h#%u%h$%zu%h%%ru%h&%ju%h'p%bu%h(`%Zu%h)P%Ru%h*@%Ju%h+0%Bu%h, %:u%h-%2u%h.%*u%h/%"u%h0%u%h1%u%h2% u%h3%u%h4%t%h5%t%h6%t%h7p%t%h8`%t%h9P%t%h:@%t%h;0%t%h< %t%h=%t%h>%t%h?%t%h@%t%hA%t%hB%t%hC%t%hD%zt%hE%rt%hF%jt%hGp%bt%hH`%Zt%hIP%Rt%hJ@%Jt%hK0%Bt%hL %:t%hM%2t%hN%*t%hO%"t%hP%t%hQ%t%hR% t%hS%t%hT%s%hU%s%hV%s%hWp%s%hX`%s%hYP%s%hZ@%s%h[0%s%h\ %s%h]%s%h^%s%h_%s%h`%s%ha%s%hb%s%hc%s%hd%zs%he%rs%hf%js%hgp%bs%hh`%Zs%hiP%Rs%hj@%Js%hk0%Bs%hl %:s%hm%2s%hn%*s%ho%"s%hp%s%hq%s%hr% s%hs%s%ht%r%hu%r%hv%r%hwp%r%hx`%r%hyP%r%hz@%r%h{0%r%h| %r%h}%r%h~%r%h%r%h%r%h%r%h%r%h%r%h%zr%h%rr%h%jr%hp%br%h`%Zr%hP%Rr%h@%Jr%h0%Br%h %:r%h%2r%h%*r%h%"r%h%r%h%r%hH&H .HHH5'HEH l%H81SHGHt HHH@`HtyHt*Hf#HtWBHtHtFHpHHuUHk%LFH5'HH81CH u HCHP01#HHuHdk%H5H8H[AUATUSQHZk%H(H]HHt|Hl%H0H9u1LmPLeXHEHHEPHEXH uHCHP0HuÃ-MtIMu IELP0MtI $u ID$LP01Z[]A\A]SHHHHtzHCHt%HCHyHH HQHHtH[ÐHSHD$HR0HD$H[fDHAWL~AVAUATUSH(MHL$LD$~oHHIIHHMHD$LÐIvLtHD$HL$HLL,LHHLHHLH\$IuH)]%HH([]A\A]A^A_fDSHH dH%(HD$1HT$Ht$HQH{HHt HCHtHH+H<$HT$Ht$.H{ HtHC H/tGH{(HtHC(H/uHGP0HCH@HD$dH3%(uH [fDHGP0"fHWHHtHB@HtH$f.AWAVAUAATAUHSHXHL$HuAAH^)HuD=F)DHDA9WHHHD;3EHkHEH=U)1Ht+HH[%1HH8ZHHt D`|HHmt?HtH+tHX[]A\A]A^A_DHCHH@0HX[]A\A]A^A_DHEHP0H|$H HItfHHHL )Lv)1111Dd$0H\$(LD$8Lt$ LL$LL$LL$L $ZI.HH+u HCHP0H0EH)u EEH=D=)DHD>LcE9~IcHHD;(LD;=)DA9A}&IcAHE9LL LT L LT uIIމ5U)E.InHE0@I.rIFLH@0fDIFLP0E@HIcHHHH)D=)5)EH|$H^HIEIHw H= DH1H1HoH)v)@h)D(HhHECHxHhH/1HGP0%fAVIAUIH= ATUSHdH%(H$1HHHH+HHLZHHHHH+IHmMtVID$M9t$ Lt6HMW%Hg H5WLH81I,$u ID$LP01H$dH3 %(H[]A\A]A^HCHP0&HmuHEHP0DHEHP0>fHCHP0HV%H H5 LH81NGfH HBMH1P11H L@f.USH8dH%(H\$(1HH|$HD$ HD$ZHHD$HH;aV%HHD$ HH;BV%HHU%HD$ H9XHU%H54H8HL.HZHHfDAVAUATUHSHHHHCHHrHUHH5u(HHHIMHHH1HXHLHIImHmu HEHP0HHPLHHH[]A\A]A^HQHHHRH5O)1HeHIuE1AeAAHmu HEHP0Mt ImH IH=kDDE11Q@HSLd$HR0HD$H[]A\A]A^IELP0AWAAHHT$6HHT$SH5Hr<1IELP0NHB@Ht_H$JAwAA+AyAA A~AAHB@Ht*H$AcAAHIHHtAVAUATUHSHHHHCHHHUHH5(HH.HIMHHH5(HH[#H5<)HLQHIImHmu HEHP0HHPLHHH[]A\A]A^fDH0HHHBH5)1HHIuE1A@AUfDHmMtImtrH H=DDE1WHSLd$HR0HD$H[]A\A]A^IELP0fAAAWwIELP0HEHP0dHHT$HHT$!H5gH 1A@AT&HB@Ht_H$A AAWA@AUAAAWAAAWHB@HtH$HH`HIDf.AUATUHSHH(L-9G%dH%(HD$1HLl$uSLFMIH=q11,H H==iLfMIHFHHD$HLl$fD&1HEHHHM(H9HWHH(L%(H=)HE(LHHHH5)1HHIH+HU H*HULe H5b(HHHIMo?HHIE1LhHL}HHI,$H+HE1HPHHUHL$dH3 %(H([]A\A]@LnH}(HD$HWR0HD$fDHCHP0HU H*H} HGP0ID$LP0H+OHCHP0@@HEHP0DfL(HH^+-AH }H=DdJ=-AHB@HH$]--AH+u HCHP0MtI,$uID$LP0낽D-Aʽ?-Aͽ-AXH1HI*H56(HvHtHD$IEMHT$LH5H(LH,}HIf.AWAVAUATUHSHHXdH%(HD$H1HHD$0HD$8LFILvL~ L;5B%HD$HD$ HD$(HlB%HL``LhhHhpMtI$MtIEHtHEH(H=a(HIHHHD$HHH5L(HHHHHD$ HT$H*HD$HHD$^I1LpH|$ HHHD$(HT$ H*]HT$HD$ H*]HD$:HHHD$\HD$(HCIML{ HD$(HD$t I,$VMt ImHt HmH;t0HHL$HdH3 %(=HX[]A\A]A^A_f.HCHP0H5U(HIHHD$0:LCH=a !H H=3@1\fH|$HGP0UH|$ HGP0H|$HGP0IELP0fHEHP0H;ID$LP0HHHD$sILxHD$H;DHXHHD$Cb!HD$ HtHHQHH)HD$(HD$ HtHHQHHHD$HD$(HtHHQHHH=(HD$gH +H=EHT$ Ht$(H|$ioHHIH==%1LpHHD$HIHL$TH) HHgIHLxLHD$ HHHL$.I.>H)HT$H*jHT$(HD$H*AHT$ HD$(H*HLLHD$ LfIKI8MMH+MIIu!H5r(HjHHD$8IMLt$0L|$8H|$HGP0#H|$ HGP0H|$(HGP0p!HQ@HH$l!nIFLP0HL$HAHP0VH|$ HGP0H|$(HGP0H|$HGP0HAHP0dg!d!HF HD$8HCHD$0!AFE11HLLHL$HD$HL$Ht%H0HVHHuH|$HL$HGP0HL$HD$ Ht%H0HVHHuH|$ HL$HGP0HL$HD$(Ht%H0HVHHuH|$(HL$HGP0HL$HtH)u HAHP0MtI.u IFLP0H H=D1BAEE11ABB!E114H !AG!AGE1!AG!AGH=A *HT$0LfH5(LHo= fAWAVIAUATUSH(HFHLnI1LHHeI~ H@$H5n(HD$HWHH&IMI~ H5D(HWHH HHH5(1HHH HQHHuHHQHHRHk9%HL@`LxhHXpMtIMtIHtHLD$GIVLH|$HD$HD$HDLD$Mt I(Mt I/!Ht H+H5(1LI,$HHH;;9%H;8%tRDHHPHHtbEHEHHPHHHUu HEHP0HH([]A\A]A^A_fH;8%tH?AHHPHHuHCHP0뒐HPHR0HSHD$HR0HD$rI@LP0HCHP0IGLP0ID$LP0H H=L81(@8H H=f1HUHIDHH H=8u1]8I,$tiHvH+lHCt$HP0t$U[8ȾY8DHB@Ht:H$HB@HtBH$8ID$t$LP0t$_II5EHff.AWAVAUATUHSHHHdH%(HD$81H6%HHD$0BLFMI@HnLs IH;-M6%LcL-(H=(LHHHHSH5(HHXHIM\H+g]HIHEHh1H5%LLHIG HHnImuI/[H;AHZ(sH{ IVHLkH5c(HHLHD$H|$YIVH5:(HHNLIMH5(1LHI/H(kHt4%HHH`HphH@pHHL$ Ht$HD$tHHD$HtHHD$HtHDHHIE1LhHH肹HIzH+Hm&IVH5(HHLIM˶HHIE1LhHL HHI,$bHmGH+-IEHIEHLHHIUuIUHD$LR0HD$H[]A\A]A^DH-9(H=z(HbHHHHSH5(HH2HIM7H+ҵHHeIE1LhHLHII,$H+HCHP0HCHP0_HSD$HR0D$fHCHP0HEHP0ID$LP0HCHP05HEHP0ID$LP0WHCHP0HmHEHP0@HHHlAvA1H "H= DDHD$HD$H0HI1DHpHH1AvA@HHHHAvA1sLAAvMtI $HQHI$t7HEHMHQHHU0HUHD$HR0HD$IT$HD$LR0HD$AvA111AvAHHP1HHgHSHD$HR0HD$NHB@HH$\LAAvLAAvHB@HH$AvA1E1hHB@HH$nAvA1HB@HH$1AvAE1AvA1AvAE1AvAAvA11BLDI^H4IH$HHIff.AWAVIAUIATIUSHHH;b"%D$HL$ L$sH(H= (HHIHIWH5(HHLHH0I//L=p(H=(L虰HH%HHSH5(HHHHHH+HT$HIHT$lIHHP Lp1H:HHHmI/H;H(sH{ H|$ HLsH5(HWHHHD$H|$MH|$ H5(HWHHRIM<H5l(1L育HIHQHIH(H %HHH`HphH@pHHL$0Ht$ HD$(tHHD$ HtHHD$(HtHݱE1HHD$8~!L$LD$ACII9uH|$8ŪH|$0tHL$0HHD$HHHHL$ HtHHD$HHH~Ht$(HtHHD$HHH|Lt$H5x(1LVHIHD$HHIkHdH;-s%H;-9%tZDHEHPHHU!EHHHHHHHEu HEHP0HHH[]A\A]A^A_fH;-%tHoAf.IGLP0HCHT$HP0HT$HPHR0IWHD$ LR0HD$ HCHP0"IGLP0HEHP0HAHP0sHAHP0AHFHP0uHEHP0H|$HGP0fD[HHA AfDL0HHE1A AHtHmt|MtHI/uBIGLP06fHHI'1A A@A AH (H= DDHHH1HEHP0uA AAH HL$HHD$HHHt>Mt.1AI/tE1!IGLE1P0AD AWHAHP0AF HB@HH$H|$ HB@Ht H$8HB@HtH$H|$ 7L裮HH|$ 葮HD$1IA AlHB@HtAH$1IA A1A H|$ 7IA HHAWAVAUIATUSHH-(H=A(H)HHMHHSH5(HHHIMH+H-(H=(HͩHH!HHSH5(HH@HIMH+=HHzH(1HLHHCLc qHH:I/H+qH}VI} H5c(HWHH>IMpI} H5>(HWHH;HH!H5(1H۫HfH HQHHHHQHHHe%HL@`LxhHXpMtIMtIHtHH}AIu@@@DAHAAIEMDL$ $t I(Mt I/Ht H+H5(1LȪI,$HHH;%H;%=DHHPHHQEH(H=W(H?HI#HIWH5(HH5LHHI/L%(H=(LHIHIWH5(HHLIMI/=SHIHE1HhL` HH荩HIH+7I/I}HmIcHI?Hc<$HI[D$ӥHH警HHe(HHZ(HPIELh Lx(L`0HX8IMHQHIUuIUH$LR0H$H[]A\A]A^A_4@4kfHARA+H;q%HAHHPHHHCHP0fDHCHP0fHCHP0JHEHP0HCHP0IGLP0fHPHR0HSH$HR0H$fIGLP0hIGLP0HEHP0 IELP0IGLP0HCHP0IGLP09I@LP0ID$LP0FHCHP0HHHA0A1fHhHHE1A0A1fDHHAfOHHAO f.HHAOf.HHIAR1AFfDLHIE1AW1AH+tbMtI/tfMtI,$t>H H=DDϷHH1AA1AID$LP0HCHP0IGLP0E1AT1AoHB@HH$`AY1ABLAA}1.LAA{1 HB@HfH$HB@HaH$LAAy1HB@HnH$A0AHB@HTH$dHB@HH$A0I,$HAE1E1\E1A0A1FE1A0A2A0A1%LAAw1(E1Ad1AA\1AA0Z1A0AA=1AID$LP01L$HÐOILHII1{LItߣHBHϣIAWAVIAUIATIUSH8H;"%HL$oH(H=(HHIHIWH5^(HHLHH,I/+L=<(H=}(LeHH!HHSH5(HHHHHH+HT$ПHIHT$hIHHP Lp1HHHHmI/H;H(sH{ H|$HLsH5(HWHHHD$H|$IH|$H5(HWHHNIM8H58(1LNHIHQHIH(H%HHH`HphH@pHHL$ Ht$HD$tHHD$HtHHD$HtH詠E1HHD$(~@LACII9uH|$(虙H|$ tHL$ HHD$(HHHHL$HtHHD$ HHHHt$HtHHD$HHHLt$H5t(1L*HIHD$HHIoHhH;-G%H;- %t^DHEHPHHU%EHHHHHHHEu HEHP0H8H[]A\A]A^A_fDH;- %tH?Af.IGLP0HCHT$HP0HT$HPHR0 IWHD$LR0HD$HCHP0&IGLP0 HEHP0HAHP0oHAHP0=HFHP0qHEHP0H|$HGP0fD+HHAAfDLHHE1AAHtHmt|MtHI/uBIGLP06fH踸HI+1AA@AsAH H=DDHHH1HEHP0uAoAAHL$HHD$HHHt>Mt.1AI/tE1!IGLE1P0AAWHAHP0AHB@HH$H|$HB@Ht H$Mt.1AI/tE1"IGLE1P0A AXHAHP0A HB@HH$H|$HB@Ht H$BHB@HtH$H|$AL蔏H#H|$肏HD$$1IAAmHB@HtAH$;IAA1AH|$(IAHHAWAVIAUIATIUSHHH;b$$HL$ L$T$~HǮ(H=(HHIHIWH5(HHLHH:I/:L=k(H=(L蔊HH0HHSH5(HHHHHH+HT$HIHT$vIHHP Lp1H5HHHmI/H;H(sH{ H|$ HLsH5 (HWHHHD$H|$WH|$ H5(HWHH\IMFH5g(1L}HIHQHIH(H$HHH`HphH@pHHL$0Ht$ HD$(tHHD$ HtHHD$(HtH؋E1HHD$8~)T$LL$$ACII9uH|$8踄H|$0tHL$0HH$HHHHL$ HtHH$HHHHt$(HtHH$HHHLt$H5>(1LLHIH$HHIrHkH;-j$H;-0$taDHEHPHHU(EHHHHHHHEu HEHP0HHH[]A\A]A^A_fH;-$tH_Af.IGLP0HCHT$HP0HT$HPHR0IWHD$ LR0HD$ HCHP0IGLP0HEHP0HAHP0nHAHP0=HFHP0oHEHP0H|$HGP0}fDKHHAHAfDL HHE1A]A HtHmt|MtHI/uBIGLP06fHأHI1AXA @A A#H H=DDHH H1HEHP0uA A#AHL$HH$HHHt>Mt.1A#I/tE1"IGLE1P0AA#XHAHP0AHB@HH$H|$ HB@Ht H$.HB@HtH$H|$ -L蔈HH|$ 肈HD$1IAbA mHB@HtAH$'IA_A 1AZH|$ (IAjHHAWIAVAUATUHSHXdH%(HD$H1HFH;$HD$0HD$8HD$@ H;A$3 H@hH H@H 1HHHH\$0q H52(HߺHD$0蔁HHHD$0 H;$H;=$DHHPHH=EHD$0 HUHBhH> H@ H1 HHHD$0ZHPL5P$L9HxHNHPH@ HT$8HD$@HHHD$@HHT$0H*HD$0H|$@HGHPHGHcЉD$ H9= T HT$@H*xHD$8HHD$@HD$8HD$H|HHUHBhHe H@ HX HHHD$08HPL9mHxH2HPH@ HT$@HD$8HHHD$8HHT$0H*HD$0Hl$@Lt$8HD$@HD$8HU$HLP`LXhL`pMtIMtIMtI$Hc(L\$LT$L(hE1AAHH|$AHHD$0LT$L\$ HD$8HHT$0H* MLl$8HD$0HD$8t I*Mt I+Mt I,$IE H8pI H5(HWHHIM I H5(HWHHHHD$@H5 (1H6HHD$0HT$@H*AHT$0HD$@H*H$HD$0HLH`LPhL@pMtIMtIMtIIIu@@@HIGL$ HEHVHEHcȉH90IGHY$I9FLLD$ LT$LL$踀LD$ LT$LL$f.IGMt I)Mt I*qMt I(JH5(1L衁I,$IML;=$L;=$DIHPHIEL%g$I$ImoHt H+H\$HtHHD$HHHVHt HmVMtI.u IFLP0HL$HdH3 %(LHX[]A\A]A^A_3@fHِ@AFtDH;=$AH|$0HHPHHH|$0HGP0fH|$0HGP0?H|$@HGP0wH|$0L\$LT$HGP0L\$LT$IELP0fHCHP0HEHP0HCHP0YH-(fWHE|HIHD$0D$Z2D$2ICLP0mIBL\$LP0L\$DID$LP0IE H8pQH5ݚ(H=^(1~HHD$@eH9HT$@H**HD$@D$3E1D$;E1HD$0HtHHqHH0^HD$8HtHHqHH0YHD$@HtHHqHH0TMt I,$Mt I/T$t$H 9H=ӠE1'MfH|$0HGP0H|$@HGP0H|$0HGP0L;=$-LS|AIHPHI&IGLP0fDI@LP0fIBLD$LP0LD$vIALD$LLT$P0LD$LT$=DH~HFHHDID$LP0EH~H^HH5(H=(1|HHD$0HHT$0H*6HD$0D$11D$.HD$E1E1E1E1ID$LP0IGLP0H|$0HGP0H|$8HGP0H|$@HGP0HHfOHHOHHWH|$@HGP0HB@H{ H$fD$1D$,1HD$H|$0HGP0D$3D$<E1E1yvHBpH5 H@H( H5(H1HHHBpH H@H H5(HD$1D$-1HD$5D$1D$-E1E1E11HD$E1H HB$H5KH8uFxH D$ /LD$(LT$ D$LL$xHD$LL$LT$ LD$(AR3D$@HD$8Ht9HHQHHu*H|$8LD$(LT$ LL$HGP0LD$(LT$ LL$HD$@HD$8Ht9HHQHHu*H|$@LD$(LT$ LL$HGP0LD$(LT$ LL$HD$0HD$@Ht9HHQHHu*H|$0LD$(LT$ LL$HGP0LD$(LT$ LL$T$H H=DLD$(LT$ LL$HD$0HT$8Ht$@H|$0LL$LT$ LD$(! LD$ LT$LL$#vHILL$LT$LD$  HD$01LLHHD$0IGHD$@HHD$@IG HD$8HHD$8IG(%xI,$HLL$LT$LD$ u3HD$ID$LLD$(LT$ LL$P0LD$(LT$ LL$HL$H7HLD$(LT$ LL$HL$贅HL$ALL$LT$ LD$(HHPHHu(HALD$ HLT$LL$P0LD$ LT$LL$EI/u(IGLD$ LLT$LL$P0LD$ LT$LL$HT$0H*u*H|$0LD$ LT$LL$HGP0LD$ LT$LL$HT$@HD$0H*u*H|$@LD$ LT$LL$HGP0LD$ LT$LL$HT$8HD$@H*u*H|$8LD$ LT$LL$HGP0LD$ LT$LL$LLLHD$8 ktHD$@Ht/HHQHHu H|$@L\$LT$HGP0L\$LT$HD$0HD$@Ht/HHQHHu H|$0L\$LT$HGP0L\$LT$HD$8HD$0Ht/HHQHHu H|$8L\$LT$HGP0L\$LT$H=(L\$LT$HD$8qLT$L\$qH tH=62[HT$@Ht$0H|$8WLT$L\$H(L(hE1AAHH|$AHILT$L\$H8uH@LP0L\$LT$HT$8H*u H|$8L\$LT$HGP0L\$LT$HT$0HD$8H*u H|$0L\$LT$HGP0L\$LT$HT$@HD$0H*u H|$@L\$LT$HGP0L\$LT$LLLHD$@KLD$ LT$LL$H qLL$LT$LD$ xD$2D$/1HD$WHB@H H$D$#3I,$D$<E1E1D$3D$<E1E1D$!3HH/$H58LD$ LT$LL$H8inLL$LT$LD$ LD$ LT$LL$qHLL$LT$LD$ YD$E2D$0E1E1E11E1HsHOsD$2D$/H;7$H!rHIxHL$0H)u H|$0HGP0ID$HD$0LLAHHD$8LAHHD$@LAվHwI,$$ID$LP0D$3D$<E1E1%H HsD$u2D$4H;D$>H.qHI HL$0H)ID$HD$0LHHHD$@LHHD$8vLվHv<I,$KID$LP0;ID$LE1E1P0D$<:D$i2D$41lD$u2D$4HD$0E11E1E1E1H$H5ȚH81pD$u2D$3D$;E1E1D$2D$/HD$H$H5hH81ypD$2D$/D$2D$9LLL{E1E1E1=asLD$ LT$LL$nD$~3LLLI{E1D$<D$z3HT$8Ht$@H|$0LD$(LT$ LL$|zHD$0HD$@HD$8D$3LL$LT$ LD$(H$HRH51H81zoHD$0D$i2D$4D$2D$7D$2D$6D$2D$4E1E11E1I,$u ID$LP0puHoD$21H|$0HGP0E1D$2D$4E1E11HxHzHPHHRHD$@HT$8CD$-2D$/E1E11HD$E1\nD$1D$.1HD${Hn$HRH51H81nHD$0D$2D$/HD$6D$m3D$i3E1vkH D$82D$/I,$u ID$LP0otD$52D$/1HbnD$52D$/vE1D$%2D$/E1E11HD$)MpHxHHPHHRHD$8HT$@-mIAH3D$?LD$ LT$LL$njHLL$LT$LD$ 7AWAVAUATUHSHHXdH%(HD$H1HHD$ HD$(HD$0HD$8LFILfLn Hn(H~0IFH9KHɁHI#A bHCHP0LnIHJcHHF0HD$8HC(HD$0HC HD$(HCHD$ H%_II IIu!H53(HScHHD$8IMLd$ Ll$(Hl$0H|$8nH H%H5H$H8fa!dH A~M#AE1I/u IGLP0MZI.PIFLP0A#A~I/ H}Hzl(uH} LeI1BLD$|$HLLDW?H/;H}t(HHL$HdH3 %(HX[]A\A]A^A_fHEHP0@IFLP0LCDH=eN*H bH=d*S1tfIFLP0YIFLP0IGLP0H}HEHP0@HCHP0I/fLD$|$HL$D5>Ha(H=k(H=HIBHIWH5_(HHLIMhI/|$@HImJ>HH#1LxHL@HHI.H+5HCHP0&fDH[HI+A=1LZHI !+A=HtH+tkH "aH=#cD1 RIGLP0IFLP0JHiZHI*A;:HCHP0LnIHfyJcHHF0HD$8HC(HD$0HC HD$(HCHD$ H7IIIIu!H5^(H;HHD$8IM Ld$ Ll$(Hl$0H|$8HHH5=mH$H8:I/ H}H*c(uH} LeI8|$LD$HALL9H1H}t&HHL$HdH3 %(HX[]A\A]A^A_HEHP0@IFLP0LCDH=[mE%"H YH=[VJ1vfIFLP0YIFLP0IGLP0H}HEHP0@HCHP0I/f|$LD$HL$A7HtX(H=a(H4HI@HIWH5Z(HHLIMfI/|$-7HIj4HH 1LxHL:7HHI.H+4HCHP0%HQHI"Ai1LQHI "AiHtH+tkH WH=YD1HIGLP0 IFLP0MHQHI"Ag[1HCHP0LnIH*pJcHHF0HD$8HC(HD$0HC HD$(HCHD$ Hu.IIIIu!H5U(H2HHD$8IM Ld$ Ll$(Hl$0H|$8HHH5bHf$H80q3HAM"AiE1I/u IGLP0MbI.XIFLP0I\"AcDHHH5@bHѢ$H8!02HXD$hHB@HH$L2M潯"AiLHB@HH$E1"AiL\2M"E1AiMgH5IU(HI0HHD$ VH5V(H0HHD$(txIH5;S(H0HHD$0t&ILP4H17H=VAj@ "6H=VA5@""Ag"Ag"Ag1HB@HtMH$e"Ag10HqH5`VM^"Ac1E1wLK3IMHT$ LUH5(LH89"i"AdL2IV0HlH5f_QH=pUA?"fDAWAVAUATUHSHHXdH%(HD$H1HHD$ HD$(HD$0HD$8LFILfLn Hn(H~00LLHD$\2HIH@HIFH9,AII.ID$H7ID$HcЉD$H9}H;-$sHQ(H=Z(H-HI9HIVH5ZS(HHaLHHI.L=8Q(H=yZ(La-HIHIVH5NR(HHFLIMvI.k-HIHE1HhL` HH 0HHH+]I/+H}5HY(uH} LeI/LD$|$HLLD,H(H}t'HHL$HdH3 %(HX[]A\A]A^A_ÐHEHP0@IFLP0MH5yP(HI,HHD$ nLCH=qR;%H PH=CRA1MIFLP0:IFLP0IGLP0H}HEHP0@HCHP0I/fLD$|$HL$Du+HO(H=GX(H/+HIBHIWH5P(HH LIMI/Hc|$-HI+HH1LxHL-HHwI.H+HCHP0DH@HHIG&A1LHHIL&AHtH+tkH bNH=PD1L?IGLP0IFLP0KHGHI)&A'HCHP0LnIHfJcHHF0HD$8HC(HD$0HC HD$(HCHD$ H%II!IIu!H5L(H3)HHD$8IMLd$ Ll$(Hl$0H|$8H"H'H5}ZH$H8F'*HAMI&AE1I/u IGLP0MdI.ZIFLP0K%AFHHs$H5ZH8&w)HD$QHB@HH$LH)MQ&ARHB@HH$E1Y&AL(MN&E1AH5L(H'HHD$(IH5I(Hi'HHD$0MI-LD*H-05&A{0&A[+&A1IHB@HtMH$.&A15'HH5&XM&A1E1LY*IMHT$ LLH5v'LHF0K%y &A#L*I H=LA(6%.H=LA5%'HsDH=HP(uH} LmIi&LD$H4$HLLL"HgH}t0HHL$HdH3 %(HX[]A\A]A^A_f.HEHP0@IFLP0"MH5IG(HI"HHD$ OLCH=TI2&H FH=&I71DIFLP01IFLP0}IGLP0H}HEHP0@HCHP0I/fLD$H4$HL$L!HE(H=O(H!HIBHIWH5F(HHULIM2I/H|$$HIY"HH1LxHL$HHI.H+ HCHP0DH?HI}'A1L>HI'AHtH+tkH 2EH=lGD16IGLP0IFLP0KHy>HI_'AHCHP0LnIfH]JcHHF0HD$8HC(HD$0HC HD$(HCHD$ HII!IIu!H5B(H HHD$8IMLd$ Ll$(Hl$0H|$8 HA'AHB@HH$0M'E1AI/u IGLP0MpI.fIFLP0WM'AE1E1'AM'AH6$H5'LH8: HH$ I~xL!H$H<$뺽4'ALIHB@HSH$yH5C(HyHHD$(=IH5@(HXHHD$0GI/M6'A1E1L9$H$B0%IHM3HT$ LDH50'LH'&#d'A1GLy!HNk'A'f'Aa'A1HB@HH$H=CAX-&L I'H=CA-&SH=CA,&(L IAWAVAUATUHSHHXdH%(HD$H1HHD$ HD$(HD$0HD$8LFILfLn Hn(H~0LLHD$< HIH@H>IFH9KfAgI.ID$HTID$HfD$H9fH;-$~Ho?(H=H(HHILHIVH55A(HH{LHHI.L=?(H=TH(L<HIHIVH51@(HH`LIMI.vHIHE1HhL` HHHHH+hI/6H}@HG(uH} LeIk|$LD$HALLOHgH}t0HHL$HdH3 %(HX[]A\A]A^A_f.HEHP0@IFLP0MH5I>(HIHHD$ }LCH=g@)$H =H=9@.1DIFLP0/IFLP0{IGLP0H}HEHP0@HCHP0I/f|$LD$HL$AH<(H=F(HHIPHIWH5=(HHLIMI/H|$HIVHH1LxHLHHI.H+ HCHP0f.H6HI%A1L5HI%AHtH+tsH *<H=w>D1-IGLP0fIFLP0>Hi5HI$AHCHP0LnIHTJcHHF0HD$8HC(HD$0HC HD$(HCHD$ HIIIIu!H59(HHHD$8IMLd$ Ll$(Hl$0H|$8nH H%H5GH$H8H AM%AE1I/u IGLP0M\I.RIFLP0C$A>H}H3$H5 HH8|7HfD$7HB@HH$oL M%ASHB@HH$E1#%ALM%E1AH5:(HKHHD$(IH57(H*HHD$0LILH@$A}$A]$A1KHB@HtMH$$A17HH5`EM$A1E1LIM{HT$ L:H5'LHM$k$ALIH=:A#{$ H=g:A#v$HuH=):A#q$AWAVAUATUHSHHXdH%(HD$H1HHD$ HD$(HD$0HD$8LFIYLfLn Hn(H~0|LLHD$HIH@HIFH9 HIHIVH57(HHJLHHI.L=5(H=>(LHIHIVH53(HH/LIM_I.LRHIHE1HhL` HHHHH+>I/ H}HZ>(uH} LeI|$LD$HALL H H}t&HHL$HdH3 %(HX[]A\A]A^A_HEHP0@IFLP0LCDH=R7 2(H 4H=$7%1vfIFLP0YIFLP0IGLP0H}HEHP0@HCHP0I/f|$LD$HL$A H3(H=<(HHI@HIWH51(HHLIMfI/|$]HIj'HH 1LxHLjHHI.H+4HCHP0%H,HI(A1L,HI (AHtH+tkH 3H=b5D1#IGLP0 IFLP0MHI,HI(A HCHP0LnIHKJcHHF0HD$8HC(HD$0HC HD$(HCHD$ H IIIIu!H50(H HHD$8IM Ld$ Ll$(Hl$0H|$8HHH5>H~$H8 HAM(AE1I/u IGLP0MbI.XIFLP0Ii(ADHHH5p=H~$H8Q HXD$hHB@HH$L M潼(ALHB@HH$E1(AL M(E1AMgH5y0(HI HHD$ VH581(H HHD$(txIH5k.(H HHD$0t&ILHaH=U2A()H= 2Ae((A(A(A1HB@HtMH$e(A1 HqH50;VMk(A1E1wL{IMHT$ LP1H5'LHh!(v(AL#I HlH5:QH=0A0(fDAWAVAUATIUHSHXH;5{$H|$8HL$ LD$H-(H=6(H HI HIUH5g/(HH LIMR ImHD-(H=6(Hm HI HIWH5.(HH| LHHBI/ HII$1HX L`HL HH I.I/H;mH 5(1LsHT$ H޿HI H8'MeH5(sH{ I9H|$H5/(HWHH{ HD$ H|$  H|$H5/(HWHH IM H55(1L$ H I4$HVHI$H0HVHHHx$HHx`HphH@pHH|$0Ht$HD$(tHHD$HtHHD$(HtHt MeE1HD$@MH\$HLIHl$8(DH(H0I8H@(HL9tdI8HH0AAI8H@I8Pt8pH(HHR8HcR H0L9uH\$HH|$@H|$0tHt$0HHD$8HHH-Ht$HtHHD$0HHHH|$(HtHHD$HHHL|$ H5&(1Lc HIHD$HHIxH H;-w$H;-Fv$DHEHPHHUEVE1HHHHHHH!Mt I.MtImu IELP0HXH[]A\A]A^A_f.1HcH4H~(H9(IHD(I8H(H)0I8HcH4H~(H;(}HH|(I8H(H0fDHP0H;0HHP0I8H0H0wDIFLP0HGP0@IELP0IGLP0JH@LP0HCHP0IGLP0jIFLP0PH@0I8H@(I8H(H+0H0HPHR0fIT$HD$LR0HD$H0HVHHfDH|$ HGP0wHFHP0fHGP0@HFHP0H5#(H=/(1kHIHI. E1H \'H=)KHAHH1$DH;-)t$HAH/(HQ E1ɋqE1 HxHD$D$$HHH8BH/(sH{ IH/(LkH|$ HIH86H|$H5x)(HWHHHD$ H|$ tH|$H5M)(HWHHHH~H5.(1HHL$HHL$PH1HVHHH0HVHHcHjr$HHp`HxhH@pHHt$0H|$HD$(tHHD$HtHHD$(HtH2E1MHD$@H\$HLIHl$8#I(IF(I0HL9tRI0HAADAFIFtA8I(HH@8Hc@ I0L9uH\$HH|$@H|$0tHt$0HHD$8HHHH|$HtHHD$0HHHHt$(HtHHD$HHHgL|$ H5v (1L<HIHD$HHIVHH;-Yq$H;-p$DHEHPHHU)EAE1fHcI0I Hy(H;(}DfHcI Hy(H;(|'IփH+(HB(uI0@I0IHH(Hx(I0HEHP0IF0I;0HIF0I0I0HHI E11DHxHI?1۽ LIܺE11[IFL E1P0DI(I+0IF0IF(I08H@HP0HPHR0HQHD$HR0HD$bH@LP0H;-0o$HAHFHP0UHFHP0HGP0@WH|$ HGP0HEHP0 E1Y E1G# Ht$ HHD$HHHuHFHL$HP0HL$Ht|E1E1HE1HHHt6MtI,$t>MI/IGT$LP0T$HAT$HP0T$ID$T$LP0T$E1 E1u E1cHB@HtUH$H|$P =! HB@Ht9H$H|$SM1۽ E1H|$HD$ H|$H E1E1 E1HB@HtH$'HB@HtH$rLlI L\HZ E1b Ht$ HHD$HHHE1E1= E1 E1 HB@Ht"H$H|$ v lH|$I P ^HB@Ht(H$H|$nHFHE1E1P0H|$`HD$ GfDAWAVAUIATUSHH8H=(dH%(HD$(1Hk$HH|$HD$ LFIItgMH= !11) [H PH= J<1HL$(dH3 %(H8[]A\A]A^A_HF HD$H^H$H$sH&( H(hE111HAHHHH8H-(H=\&(HDHIXHIT$H5(HHLHHI,$L=(H=%(LHIJHIT$H5Z(HHgLIMlI,$NfWeHI?HIYH1L` HXHLzHIDI.I/HIL`1HH3HIUHmI/L;%Si$L;%h$I$HPHI$Me HT$HH5g$I$MI}HHYI,$H+u HCHP0HiHh$HHD$@Hh$HD$pfWf.Me HT$H52g$I$LI}dHH=I,$ID$LP0of.L;% h$L|\AE11I$E1HHI$ID$4$LP04$H@HP0ID$LP0TID$LP0IGLP0IFLP0IGLP0HEHP0ID$LP02ID$LP0eHHI[\AH H=D1 H$/f.LXHI`\AE1E1Ht HmMt I.MtI/uIG4$LP04$l@H5 (H=!(13HHdHAHE\HHHEHEHP0\ HE4$HP04$[IF4$LP04$UH5(H=+!(1HHHH+t\A1HCHA1P0\oHnHtxHtiHIHHI(Hu&M~*H5(HHkHD$IMYHD$H\$HD$HF HD$HFHD$돾x\AE14L\AHB@H1H$a]\AE1s\AE1HB@HH$b\A\AE1yg\Aio\Ae\AE1|HH5(HHHD$I1۾2\AE1\A\A1HT$LXH5'HH>[=L H1LIAWAVAUIATUHSHH8dH%(HD$(1HBc$HHD$HD$,LFIIHF HD$H[H$H$H( H(hE111HAHHHH8H-(H=_(HGHI{HIVH5(HHLHHRI.AL=(H=(LHIGHIVH5`(HHLIMI.fWlHIFHIHH1Lp HXHLHIYI,$I/HILp1HH9HI!HmI/L;5Ya$L;5`$dIHPHIvVMu HT$HH5C`$IMI}HHmI.EH+u HCHP0H_H51(HI]HHD$LCH=@1@UH gH= " S1HL$(dH3 %(9H8[]A\A]A^A_fDH!`$HD$ fWf.Mu HT$H5J_$ILI}\HH/I.IFLP0DL;5_$LCeVA E11IE1HHIuIF4$LP04$Ht Hm"Mt I,$MI/IG4$LP04$H@HP0lfIFLP0IFLP0ID$LP0I/KIGLP0<@IGLP0qfHEHP0OIFLP0{IFLP0H5j (H=(1$H$aHj( H(hE111HAHHHH8H- (H=(HHI{HIVH5(HHLHHRI.AL=r (H=(LHIGHIVH5 (HHLIMI.fWHIHIHH1Lp HXHL1HIYI,$I/HILp1HHHI!HmI/L;5 Y$L;5W$dIHPHIvVMu HT$HH5W$IMI};HHmI.EH+u HCHP0H_H5 (HI HHD$LCH= 1TH  H=  1HL$(dH3 %(9H8[]A\A]A^A_fDHW$HD$ fWf.Mu HT$H5V$ILI}2THH/I.IFLP0DL;5aW$LIUA E11IE1HHIuIF4$LP04$Ht Hm"Mt I,$MI/IG4$LP04$H@HP0lfIFLP0IFLP0ID$LP0I/KIGLP0<@IGLP0qfHEHP0OIFLP0{IFLP0H5*(H=k(1HHHHH+TA 1ېH H= D1HJUf.H8HIu(UA LHI-UA E1E1GfDH5Y(H=(1#HH%HA yHEXUHHHE/HEHP0XUID$4$LP04$HE4$HP04$HCHA 1P0TLvItxItiMM[HMIIu&M~*H5(H%HHD$IMHD$H\$HD$HF HD$HCHD$돾4UA oUA E1@UA E1$dIHPHIvVMu HT$HH5?$IMI}KHHmI.EH+u HCHP0H_H5'HIHHD$LCH=l1NH 'H=5 1HL$(dH3 %(9H8[]A\A]A^A_fDH>$HD$ fWf.Mu HT$H5>$ILI}B;HH/I.IFLP0DL;5q>$LNA E11IE1HHIuIF4$LP04$Ht Hm"Mt I,$MI/IG4$LP04$H@HP0lfIFLP0IFLP0ID$LP0I/KIGLP0<@IGLP0qfHEHP0OIFLP0{IFLP0H5'H={'1HHHXH+VNA 1ېH H=D1HJUf.HHHIuNA L HINA E1E1GfDH5'H='13HH%HA HENHHHE/HEHP0NID$4$LP04$HE4$HP04$HCHA 1P0VNLvItxItiMM[HMIIu&M~*H5'H5HHD$IMHD$H\$HD$HF HD$HCHD$돾NA NA E1NA E1NA NA E1NA NA pNA E1]HB@HH$HB@HH$^NA E1@1۾mNA E1 HT$LqH5'LHNNA "RNA 1LxI8LhHAWAVAUIATUHSHH8dH%(HD$(1H9$HHD$HD$,LFIIHF HD$H[H$TH$!H*' H(hE111HAHHHH8H-'H='HHI{HIVH5L'HHLHHRI.AL=2'H=s'L[HIGHIVH5'HHLIMI.fWHIHIHH1Lp HXHLHIYI,$I/fHILp1HHHI!HmI/L;57$L;56$dIHPHIvVMu HT$HH56$IMI}HHmI.EH+u HCHP0H_H5'HIHHD$LCH=&1GH H=1HL$(dH3 %(9H8[]A\A]A^A_fDH6$HD$ fWf.Mu HT$H55$ILI}2HH/I.IFLP0DL;5!6$L^HAcE11IE1HHIuIF4$LP04$Ht Hm"Mt I,$MI/IG4$LP04$H@HP0lfIFLP0IFLP0ID$LP0I/KIGLP0<@IGLP0qfHEHP0OIFLP0{IFLP0H5'H=+'1HHHH+GA\1ېH H=zD1HJUf.HHIu=HAcLHIBHAcE1E1GfDH5'H=b'1HH%HAd9HEmHHHHE/HEHP0mHID$4$LP04$HE4$HP04$HCHA\1P0G_LvItxItiMM[H諽MIIu&M~*H5u'HHHD$IMHD$H\$HD$HF HD$HCHD$돾IHAcHAeE1UHAcE1QHAcZHAcE1.HAaDHAcp?HAcE1]HB@HH$HB@HH$^GHAcE1@1۾HA]E1 HT$L+H5'LHGiHAd"GA\1L(I8LHAWAVAUIATUSHH8H='dH%(HD$(1H^1$HH|$HD$ LFIItgMH=u11/FH H=1HL$(dH3 %(H8[]A\A]A^A_HF HD$H^HK$H$snHw' H(hE111HAHHHH8H-'H='HHIXHIT$H5'HHLHHI,$L=}'H='L覾HIJHIT$H5'HHgLIMlI,$NfW%HIHIYH1L` HXHL:HIDI.I/谾HIL`1HHHIUHmI/L;%/$L;%-$I$HPHI$Me HT$HH53/$I$MI}BHHYI,$H+u HCHP0HiHY.$HHD$@HA.$HD$pfWf.Me HT$H5.$I$LI}*HH=I,$ID$LP0of.L;%-$L[FAE11I$E1HHI$ID$4$LP04$H@HP0ID$LP0TID$LP0IGLP0IFLP0IGLP0HEHP0ID$LP02ID$LP0eHhHIFAH H=D1H$/f.LHIFAE1E1Ht HmMt I.MtI/uIG4$LP04$l@H5)'H=r'1HHdHAIHEFHHHEHEHP0F HE4$HP04$[IF4$LP04$UH5'H='1lHHHH+tqFA1HCHA1P0qF/HnHtxHtiHIH{HI(Hu&M~*H5E'H赹HkHD$IMYHD$H\$HD$HF HD$HFHD$돾FAE14FAHB@H1H$aFAE1FAE1HB@HH$FAFAE1yFAiFAFAE1|HH5_'H菸HHD$I1۾FAE1FAmFA1HT$LH5'HH F=LɻH1L蹻IAWAVAUIATUHSHHdH%(H$1H($HHD$pHD$xH$k LFItNItfDAWAVAUIATUHSHH8dH%(HD$(1H$HHD$HD$ LFII3HF HD$H[HN$褝H$~qHz' H(hE111HAHHH H8mH-޿'H='HHIHIWH5'HHJ LHH| I/!L5'H='L諛HIHIWH5 'HH LIM I/fW,HIb HI_ H1Lx HXHLAHIJ I,$I.趛HI Lx1HHHI HmjI.tL;= $L;= $IHPHIVH-?'H='HhHI<HIWH5'HH)LIMI/H-'H=$'H HIHIWH59'HHiLHHnI/艙HIAcHI>H1Lx HXHH螜HIHmwI,$\HI#Lx1HLUHI+I.?I,$IL;=u $L;=; $IHPHI* M} HT$HH5$IMI}gHII/H+u HCHP0LcH5'HIyHHD$LCH=1XjH H=ؿk1HL$(dH3 %( H8[]A\A]A^A_fDH9 $HD$efWf.If.dM} HT$H5$ILI}HII/IGLP0L;=$LKkAE1E11I/uIG4$LP04$Ht HmMt I,$MI.IF4$LP04$L;=)$L軘kAE1E11kfH@HP0IGLP0IGLP0ID$LP0I.kIFLP0\@HEHP0I.IFLP0}DIGLP0IGLP0IGLP0;ID$LP0HEHP0zIFLP0I,$ID$LP0IGLP0fIGLP0H5"'H='1蜘HHHH+87kA1ېH H=DE1wHfHHIkALHIIkAE1E1fDH5I'H=R'1ӗHHFHA)HEkHHHE/HEHP0kH5'H='1sHHHϮH+QkA1HHIkA@HHIkA@H5I'H=Z'1ۖHHHA1HEkHHHE7HEHP0k#HE4$HP04$3ID$4$LP04$*HCHA1P0QkHCHA1P07k5LvItxItiMMH聎MIIIu&M~*H5K'H軒HHD$IMHD$H\$HD$HF HD$HCHD$HB@HH$E1kAE1侓kAE1kAkAE11HB@HH$kAkAE1kA}kAkAE1ZHB@HZH$HB@HH$kA4kAE1!kAkAE1kAE11kA1lAE111۾hkAE1kAHT$LH52f'LH蒚jbL]H3kA13MkA1!L)HkAL I LIAWAVAUIATUHSHH8dH%(HD$(1HB$HHD$HD$ LFII3HF HD$H[H莑$H$~豏H' H(hE111HAHHH H8mH-'H=_'HGHIHIWH5ܵ'HHJ LHH| I/!L5²'H='LHIHIWH5h'HH LIM I/fWlHIb FHI_ H1Lx HXHL聑HIJ I,$I.HI Lx1HH9HI HmjI.tL;=Y#L;=#IHPHIVH-'H='H訍HI<HIWH5='HH)LIMI/H-#'H=d'HLHIHIWH5'HHiLHHnI/ɌHIA裍HI>H1Lx HXHHޏHIHmwI,$\RHI#Lx1HL蕏HI+I.?I,$IL;=#L;={#IHPHI* M} HT$HH5#IMI}HII/H+u HCHP0LcH5]'HI蹋HHD$LCH=l1蘛gH H=5諠1HL$(dH3 %( H8[]A\A]A^A_fDHy#HD$efWf.If.dM} HT$H5#ILI} HII/IGLP0L;=#L苌gAE1E11I/uIG4$LP04$Ht HmMt I,$MI.IF4$LP04$L;=i#L hAE1E11kfH@HP0IGLP0IGLP0ID$LP0I.kIFLP0\@HEHP0I.IFLP0}DIGLP0IGLP0IGLP0;ID$LP0HEHP0zIFLP0I,$ID$LP0IGLP0fIGLP0H5¤'H=['1܋HHH8H+8TgA1ېH ̬H=BDE1距HfH(HIgALHIIgAE1E1fDH5'H='1HHFHAiHEgHHHE/HEHP0gH5'H=2'1賊HHHH+ngA1H(HIgA@HHIgA@H5'H='1HHHAqHEhHHHE7HEHP0h#HE4$HP04$3ID$4$LP04$*HCHA1P0ngHCHA1P0TguLvItxItiMMHMIIIu&M~*H5'HHHD$IMHD$H\$HD$HF HD$HCHD$HB@HH$E1gAE1侰gAE1gAgAE11HB@HH$hAgAE1gA}gAgAE1ZHB@HZH$HB@HH$gA4gAE1!gAgAE1gAE11hA11hAE111۾gAE1hAHT$LܫH5"Y'LHҍgbL蝇HPgA13jgA1!LiHgALII L9IAWAVAUIATUHSHH8dH%(HD$(1H#HHD$HD$,LFIIHF HD$H[H΄$$H$H' H(hE111HAHHHH8H-^'H='H臂HI{HIVH5'HHLHHRI.AL='H=C'L+HIGHIVH5'HHLIMI.Ϳ訁HI肂HIDH1Lp HXHL轄HIUI,$I/2HILp1HHuHIHmI/|L;5#L;5[#hIHPHIrRMu HT$HH5g#IMI}HHiI.AH+u HCHP0HcH5m'HI虀HHD$LCH=i1xeH H=2l苕1HL$(dH3 %(1H8[]A\A]A^A_fDHY#HD$ f.Mu HT$H5f#ILI}HH#I.IFLP0L;5#L{|fAE11IE1HHIuIF4$LP04$Ht HmMt I,$MI/IG4$LP04$H@HP0dIFLP0IFLP0ID$LP0I/OIGLP0@@IGLP0ufHEHP0SIFLP0IFLP0H5r'H='1|HHHؘH+@fA1ېH lH=D1XHNYf.HțHIufAL蠛HIfAE1E1OfDH5'H=2'1賀HH%HA HEfHHHE/HEHP0fID$4$LP04$HE4$HP04$HCHA1P0@f/{LvItxItiMMcH{xMI#Iu&M~*H5E'H|HHD$IMHD$H\$HD$HF HD$HCHD$돾fAfAE1fAE1fAfAE1qfAfAxfAE1eHB@HH$HB@HH$^fAE1H1۾WfAE1HT$LvH5P'LHOefA"H-'H='HyHIHIUH5m'HH LHH ImL5R'H='L{yHIHIUH5'HH LIM ImyHI HHX1H'LLHIE |HI I.dImnyHI Lx1HH{HI HmCImMH#H #I9H$HL$I9UIHPHIH-'H=I'H1xHIHIWH5ƞ'HH- LIM2 I/H-'H='HwHIHIWH5 'HHO LHHT I/IT$H5'HHQ LIMf xHIc H1Lx HXHHSzHIt Hm,I.wHILx1HL zHIImI.L;<$L;|$IHPHIM|$ HT$HH5#IMI|$ HII/H+u HCHP0LfHy#HD$Ha#HD$L;=I#LwAIeeLL;=# LwAegH@HP0IELP0IELP0MIFLP0ImIELP0@HEHP0ImIELP0@IGLP0IGLP0&IGLP0rIFLP0HEHP0IELP0I.IFLP0DIGLP0 fIGLP07H5!'1HrHH1H#H #H9H$HL$H9DHEHPHHUEIT$H5'HHLHHHHqHIHmBL;,$L;l$IEHPHIU#kMl$ HT$H5#D$IELI|$蒵HIImIELP0@HHIlA*eefH \H=DE1GHfL踐HIqA/eeE1E1HmuHE$HP0$Mt ImMt I.MlI/bIG$LP0$MDH5y'H='1suHIWHAXeɌIfHHIIGLP0fDHЏHIAdeg@H訏HIAiegE1fH5'H=:'1tHI4HAeIhHHIIIGLP0h5DH;-#HsAAd[E1E1E11JDL;-I#HLr9Ad]E1E11%fIF$LP0$&IE$LP0$HUHR0bHEHP0IELP0H5'H='1|sHHH؊H+tTAd\1H5I'H='1;sHIH藊Imt.Ad^1HCHAd1P0\IELAd1P0^mH^H:H'HIH%kHIHu&M~*H5'H_oHvHD$(IMdHD$(H\$ HD$A@eeE1RAHH1LxHL@HHD$8<HmQ H|$H5`'HWHHIMIGL9 H;#H@hHH@H1LHHI/ Ht$8HV HFHHH9H9HEHVH,HHFHm HSH5_'HHHHHG=HIHD$81LHHIG?HInHmj I/P IUH55_'HH(LIMIm H|$H5^'HWHHIM`IEL9 H;# H@hH}H@Hp1LHHQIm?HHI/}Im;HH-HD$H|$@1HHHE >HIHmQIEH;6#I}HIEHD$HIE HD$(IE(HD$0HD$HHHD$(HHD$0HImL5J^'H=g'Ls:HI HIUH5\'HHLIMImH|$HH5a'HWHHIMHt$0L7HIImz:HILx;HI:H8['H5)`'H<Y LLLHB@HqH$uHD$(HD$8FmQHB@HJH$H|$HD$(HD$8mHD$LMLMLR#IxnpE1E11HBHHtHH$TnuE1LnE1tE1nsE1E1LH LILH5HIH|$qIdHR#H5GH81 nL?IH|$-IH&#HRH5[JH81m[Ld$0L|$(3nlE1HD$PHD$0HD$(HD$HAI.u IFLP0WuLSL|$(;nlE1E1AE1HD$(IELP0)nl|I}HLIEHHL$HHHH@HL$(HD$0HIH|$IH=VA1A')VlLIH|$IHHAHT$`L@H5h&LL"_lXH|$HaIH|$OIH|$=IHD$(HD$8mUH IH|$ILILHRH|$IGAWAVIAUATUHSHHdH%(H$1H #HHD$`HD$hHD$pH8'HD$xLFI~4IIHF0HD$ HC(HD$Hk &IHD$ H#HD$HH[HD$HHD$PHD$XHHEH;-S#] L%:'H=D'LHI HIWH5<'HH2LHHD$HI/eHIHD$ 1LHIGH|$HHHD$PrHT$HH*HD$HI/H|$PH5Q:'HWHHIMHT$PH*L%<'H=C'HD$PLHI} HLLJI7HVHIc L%<'H=B'LHI HLL1HHD$PI/yHD$PHPH;#HxH"HP L`H@(HT$HHD$XI$HD$HHHD$XHHT$PH*)HD$PHD$H1HLHD$HH$HD$XHD$XHD$HHHD$PH;#H;=#QDHHPHHoEHD$P H4$HgHHHD$XH;#H;=}#:DHHPHHXEHD$XI HߺHHHHD$PH;Q#H;=#DHHPHHEHD$P I~ H5:'HWHHHD$H|$I~ H5:'HWHHo HHD$PH5C@'1HYHHD$XHT$PH*HT$XHD$PH*H݃#HD$XHHH`HphH@pHHL$8Ht$(HD$0tHHD$(HtHHD$0HtHH5:4'HHHD$XJHHD$PHHL$HD$PHT$XH|$HP 1HXHHH(IF(HHt$PIF(HD$XHF0OHHD$XhHT$PH*HD$H;F#Lt$XHD$PHD$XI ILt$HT$0Ht$(H|$8 H50'H|$1HL$IHHD$ HHHAML;=#L;=#|I7HVHIn ImMt I,$H $HtHHD$HHHHL$HtHH$HHHMt I.Ht H+;HtHmu HEHP0HD$ELCf.H=81 5H 4H=8%1H$dH3%(HĘ[]A\A]A^A_H;=#AH|$PHHPHHH|$PHWR0f.H;=9#AH|$XHHPHHH|$XHWR0f.IGLP0H|$HHGP0HD$HI/4IGLP0%fH|$PHGP0OIW$LR0$H;=y#_A H|$PHHPHHNH|$PHWR0=f.IGLP0xH|$PHGP0HD$ GfDHD$ H#HD$4fHCHP0IELP06HAHP0WHAHP0hIFLP0gID$LP0H|$XHGP0+H|$PHGP0H|$PHGP0L;=1#wLI7HVHIrIWD$LR0D$[fH~#HH ~#HH$HHHH/'HH3HVHHHHQfDH|$HGP0L+HIEE1E1D$6D$E1HD$H$f1HD$HHtH0HNHHHD$PHt H(HD$XHtH0HNHH@Ht H*QT$t$H 0H=4!MHD$$L0*HIsE1D$6D$E1HD$H$(H=a4'L HI HHD$PTH=v8'1LxHHI HT$PH*LHD$P<&I,$D$76D$E1HD$H$E1uDH|$XHT$ HGP0HT$ DHBHP0H|$HHT$ HGP0HT$ DDH|$PHT$ HGP0HT$ 8DL(HI`E1D$C6D$E1HD$H$v HHD$PIEHt$PH=2'Ln~ HHD$XcHT$PH*1HD$P HHD$PHT$XH=6'HHD$XHP1M HHD$XHT$PH*H|$XHD$P$HT$XH*~HD$XD$6E1D$D~ HHD$XKIEHt$XH=1'Ln HHD$P HT$XH*lHD$X& HHD$X}HT$PH=5'HHD$PHP1U HHD$PHT$XH*H|$PHD$X#HT$PH*HD$PD$6E1D$DH5''H=Z5'1 HHD$PH5#HT$PH*HD$PD$6E1D$uDH|$XHGP0qH|$PHGP07H|$PHGP0H|$PHGP0(H|$XHGP0H|$XHGP0H|$PHGP0UHB@HH${DHCHP0HW)'Qf.HAHP0HD$ H=4'HHD$XL=L+'L|HnHHD$PHHH5-'HH HHHD$HHT$PH*Ht$HH|$XHD$PHHD$PHT$HH*H|$PH;=Ox#HD$HH;= w#BDH/\EHD$PlHT$XH*LEHD$X>HHD$X IH|$ 1Ht$XLvw HHD$PW HT$XH*HD$PHD$XHD$PHD$9fH|$PHGP0BID$LP0GD$7D$D$E6D$1E1HD$H$E1I/IGHT$ LP0HT$ H;=v#UA H|$PH|$PHGP0H|$XHGP0L=('H=?2'L'HHHHD$PHWH5*'HH HHD$H HT$PH*XHt$HH|$XHD$PHHD$P HT$HH*H|$PH;=u#HD$HH;=t##DH/EHD$PL=('H=V1'L>HH2HHD$PHWH5~('HH HHD$H HT$PH*Ht$HH|$XHD$PHHD$Pg HT$HH*WH|$PH;=u#HD$HH;=s#tfDH/EHD$PH|$PHGP05H|$HHGP0^H;=t#AH|$PH;=bt#tAH|$PzL HHD$PA7AHD$HHt#HHQHHuH|$Ht$HGP0t$HD$XHD$HHt#HHQHHuH|$Xt$HGP0t$HD$PHD$XHt#HHQHHuH|$Pt$HGP0t$H &H=y*DHD$PlHT$HHt$XH|$PhHI@ HD$PH|$1LHHD$PIGHD$XHHD$XIG HD$HHHD$HIG(Ht$HHHD$HHHuHFHL$HP0HL$H HHL$HL$H1HVHHuHQD$HR0D$m + I/u IGLP0HT$PH*u H|$PHGP0HT$XHD$PH*u H|$XHGP0HT$HHD$XH*u H|$HHGP0HT$0Ht$(H|$8HD$HwHq#HHD$LH|$HHGP0H|$PHGP0H|$PHGP0H|$XHGP0$H|$HHGP0H|$PHGP0MH|$PHGP0LHHHD$PK7AfLHHHD$PY7ALnI"HA=JcHHF0HD$xHC(HD$pHC HD$hHCHD$`H(IIIIu&M~*H5!%'HQHiHD$xIMWHD$pHl$hH\$`HD$HD$xHD$ HB@HWH$D$ 6D$1E1"E1E1D$6D$E1HD$H$D$6D$1E1HB@H7H$D$6D$1E1D$6D$E1HD$H$AH;p#HHHHL$PH)uH|$PH$HGP0H$HBHH$HD$PLAHIH$HAHHD$HH$HAHHD$XH$HA׾H@H$H*zHBHP0kD$6D$E1SD$6D$E1;D$6D$E1#HHxD$R6D$E1HD$H$E1MH5+!'HIHHD$`MzH5!'HHtHD$hIMTH5 'H{HHD$pID$7HL$HHD$ HHH1D$E1/D$6D$E1D$6D$E1D$6D$E1D$7D$E1HB@H H$-(7AE17AE1D$7*7AE1Hl#H5(H81BD$7E1HT$0Ht$(H|$8 LD$MMqM1IdHAHE1P0D$D$7D$HT$`L"#H5l&LH5jIzMHQ@Ht'H$9Hk`7AH"^7Ai[7AhHR7A;HB@Ht"H$M7HB@HtH$DD$6D$E1~atHD$%IMD$w6D$AH*u HBHP0uLD$6D$E1HD$H$1AE1D$m6HxHH@L HPH@HT$HHD$Xw7Ar7AP7AD$26D$E1HD$H$D$-6D$1E1D$+6D$E1HD$H$VLtHH7AFC7F7A,HT$HHt$XH|$P,HD$PHD$XHD$HD$7D$7D$7tD$7gD$6D$E1D$6D$E1}D$6D$E1eD$6D$E1MD$6D$E15D$6D$E1D$6D$E1D$6D$E1f.AWAVIAUIATULSHHXH;5g#H|$ LD$(L%p'H=#'LHI HIWH56'HH@ LHHA I/.HL$0HIHL$0 HHX1Hg#HLHID$ .HHHL$0 H)(I,$ H;L "'1LcLD$(LLH޿AHI H8MoH"'sH{ I9HUH5'HH HIMs HUH5'HH9 HIM H5!"'1L7HHc I.Hm;He#HHx`HphH@pHH|$(Ht$0HD$8tHHD$0HtHHD$8HtHIoE1HD$@HH\$HH\$ LMIIHHH0I@H0I8H H0HH0MKAGE1IG -yH(H0H0H@(AE;O}NIcI4H0H@H0Pt8NH(AHR8HcR H0E;O|II9)H\$HMH|$@H|$(tH|$(HHD$ HHHH|$0HtHHD$ HHHH|$8HtHHD$ HHHH5'1L5ImHUH H;-_d#AH;-$c#DHmEEHIHHHHHMtI/u IGLP0HXL[]A\A]A^A_@iHcHH8u H@HP0HCHUH5'HD$(HHHIMHUH5'HHHHHH5'1HHL$0HHHL$0H)u HAHP0Hmu HEHP0H_#HLp`LhhHhpMtIMtIEHtHEHD$8IGE1HHD$0HH\$@LLIH@H|$ H0H8H0H0H H0HH07H|$(E1HߋUHE +xDH(H0H0H@(AD;U}OIcHtH0H@H0Pt8 H(AHR8HcR H0D;U|HH;\$0"H\$@LIHH|$8~MtI.u IFLP0MtImu IELP0HtHmu HEHP0H5 '1LE1M|HB@HtH$WE1zAHH=H6 'AvI~ IUHMfH5>'HHILHD$H|$!IUH5'HHLIMH5 '1LHIHQHIwHHQHHTH=N#HHH`HphH@pHHL$8Ht$ HD$0tHHD$ HtHHD$0HtHE1HHD$@~RLt$HLILt$(Ll$(LLl$IH|$LLHIHL9uLt$HLl$(H|$@H|$8tHL$8HHD$HHHHL$ HtHHD$HHHHL$0HtHHD$HHHSL|$H5&1LMHIHD$HHI2 HH;jM#H;0L#HHPHH ILHIHHHu HCHP0MImIELP0fL;%L#sLKAdiwE1fDHAHP0_fHAHP0%H;-YL#HAiyE1E1MfDL;5!L#LA.j{E1E11fDMt I.Ht HmYMI,$ ID$T$ Lt$P0T$ t$DHH@P0fHH@P0(H@HP0IELP0IELP0HEHP0LL&1LHt$0AHIH8u H@LP0H&Aw(E1E1IW0 HxHD$D$$HHH8u H@HP0HCHUH5y&HD$8HH9HIMHUH5P&HHHIMH5&1LHHxImu IELP0Hmu HEHP0Hy #HLh`HhhLppMtIEHtHEMtIXHD$@IGE1HHD$0H\$HLMILH@H|$ H0H8H0H0H0 HD$(HD$8E1SHC &rH(H0H0H@(AD9S~NIcH4H0H@H0Pt8H(AHR8HcR H0D9SHH;l$0"LMIH\$HH|$@WMtImu IELP0HtHmu HEHP0MtI.u IFLP0H5&1LI,$Hu ID$LP0H|H;-8#AH;-#DyEHmu HEHP0E0A3A7Xf.tuy9fHD(H0H(H)0H0HcHfDL;5#LSA`?E1E1MtI.uIFT$Lt$P0T$t$ffDHH@P0tH@HP0H@HP0ID$LP0+ID$LP0yID$LP0IGLP0ID$LP0IFLP0IELP0IELP0XIELP0IELP0IFLP0ID$LP0Im IELP0IFLP0fIGLP0kIGLP0IGLP0ID$LP0IGLP0)IELP0IFLP0HHIZIFLP0_>AL8HI_?DLHI_?DH5&H=&1+HIiH臶I$`@HHIIFLP0$`@H5d&H=E&1ƞHHH"H+tg$_211HD$@H5!&H=&1{HHwH׵H+t7 _011HD$HCHP0HCHP0HCHP0)LvIHJcHHF0HD$HID$(HD$@ID$ HD$8ID$HD$0HLIHtbItwIu&H~*H5 &HzHHD$HHHHD$HLd$0H\$8Hl$@HD$(H5z&H2HHD$8fHH5&HHHD$@Hl1U_71HB@HH$_=r_;E1`_;E1HB@HH$E1_=z_=E1hHB@H}H$_;E1E1_=._=_=E1 E1_;E1s_9HB@HNH$_;E1d_81|_;E1M;`AE15_?E1`?h`?Y `?`?wHB@HH$E11HD$;_3E11HB@HkH$_?L蒛I_HT$0LH5l&LH臡E^/LRILBI_<;_>,MH=21AF|^LILԚIs _21HD$^.1HD$_01HD$H=1A谦w^+ `@WL?I:AWAVAUATIUSHHhdH%(HD$X1H#HH|$HD$0HD$8HD$@HD$HoLFIIHF0HD$I\$(Il$ Md$L诗HD$街HD$ 蓗D$(H躕H& L(hE111AHLAHHD$H8cH|& L(hE111AHHAHIIH8.H7& H(hE111AHHHHH8H&H=&HĔHI HIUH5Y&HHLHHhImL5>&H=&LgHI HIUH5ܸ&HH1LIM Im`HI 轔HIHD$1LLHMo IGHII.I/iHIY Lh1HH謖HI H+I/L;-#L;-#IEHPHIUH&H=2&HHI& HIUH5&HH LIM ImH&H=տ&H轒HIHIUH52&HH LHH Im6fW=HI HI I$1Lh L`HHQHI H+I.ǒHI Lh1HL HI I/I.L;-+#L;-#IEHPHIUHO&H=&HxHI HIUH5 &HH LIM ImaH&H=3&HHIHIUH5&HH LIM ImfW蛐HI uHH HE1Lh HhHL诓HI I/ H+%HH Lh1HLhHI I.H+L;-#L;-O#IEHPHIUHD$HL$IHT$H5s#MLh IEHxL,$HH ImSHL$HHD$HHHMt I,$nHtHmu HEHP0HBMD$fDH=91HLH oH=0 [1HL$XdH3 %(XHh[]A\A]A^A_fDH)#HD$f.D$fWf.D$ f.D$(7HD$HT$H57#T$(L$ Lh D$IELHx0HH ImIELP0f.L;-y"L !M E1E11^DID$LP0HAHP0[L;-"%L諏]M E1E11ImuIEt$LT$P0t$T$Ht H+Mt I.MwI/mIGt$LT$P0t$T$NfDL;-q"LM E1E11SfHH@P0H@LP0H@HP0IELP0DIELP0IGLP0IFLP0HCHP0I/IGLP0DIELP0!fIELP0nIELP0HCHP0I. IFLP0DIGLP0I.*IFLP0DIELP0;IELP0IELP0IFLP0H+tHCHP0eDHCHP0 fIGLP0IELP0]IELP0H5&H=˷&1LHIH訥I,$CnLy 1E1HD$DH ,H=½1H|$,Df.H舨HIM LhHI5M E1E1H51&H=&1{HHHפH0M HHH:HCHP00M !HاHIOAM }DH谧HIM E11(1L E1MuH5&HH諆HHD$0BH5~&H莆HHD$8HH5U&HmHHD$@H M E1M E1M E1qM E1E1\HB@HH$HB@HH$FL 1L M E1HB@HhH$FM E1HM PM CM E1YM TM 1HB@HH$0 M aM sM 1AM E1E1,M M E1 M E1M 1HD$L~ E1E1E1zM E11HB@HqH$\HB@HkH$M M (LЇILHgMHT$0LԬH5=V&LH譍L,M hM L} 1HD$H=q1AzK-H=I1ARKLHLI1jLy 1HD$L{ 1HD$L蜆IL茆I@AWAVAUMATIUSHhH;"H|$8Ht$@HL$ LD$(LHL&H=&HuHI HIWH5&HHLHHI//H&H=1&HHI HIVH5~&HHLHHI.艂HII$1HX L`HHÄHHHmI.H;L&1LcHL$(HT$ H޿AHI[H8IoHQ&sH{ H9@IUH5X&HH LHD$ H|$  IUH5/&HH LIM@ H5&1L̓H$I}HWHIUSH8HWHH1HU"HHx`HphLhpHH|$0Ht$(tHHD$(HtHMtIE&IoE1HD$HH6H\$PLl$XMH\$@MLt$8dDH(H0I8HB(I@HBI@JH(H0I@HB(II9I@LH0I8H0 CI8HBI8JT8H(HI8HcI H0F@8H(IHI8HcI H0I9WMH\$PLl$XH|$HzH|$0tH|$0HHD$8HHHH|$(HtHHD$0HHHMt ImLt$ H5&1L蔁HIHD$(HHIFH H;-"H;-w"DHEHPHHUE' HHPHHHHHMt#IHSHIuIWHD$ LR0HD$ Hh[]A\A]A^A_fDTHcH]/]E1HB@HjH$HB@HdH$]vHB@HOH$E1]HB@H7H$E1]]E1M]E1E1T]YE11E1|]E1?MnM2^E1]^H=1At\oHT$ LɍH58&LHon\?L:hI1L*hH}LhIL hHaa]1E1G]1E1f.AWAVAUATUSHHHH 8"dH%(HD$81H<$H]&HH=K&HL$0HD$ H|$(LFI8~.II HF(HD$H^ Hn"MHH"HHD$H-eHD$eD$tdH FcHO& L(hE111AHHAHIjH;4"H& H(hE111HAHHHH;"H-h&H=&HbHIHIVH5&&HHjLHHI.L- &H=M&L5bHIYHIVH5&HHLIMI.fWaHIbbHI=H1Lp HXHLdHI I/uImZ@bHILp1HHdHIHmLIm1L;5"L;5h"eIHPHIH$HT$IH5"LHh HEIHxHIHmI,$ZHtH+u HCHP0LDMM@H=ĉ11pZH H= u1HL$8dH3 %(7HH[]A\A]A^A_ffWf.D$H$HT$H58"L$D$Hh HEHHx诮HIHm3HEHP0$fL;5"La\[EnfH"HD$H"HHD$@ID$LP0IFLP0IFLP0BIELP0IGLP0|IELP0HEHP0IFLP0HEHP0 L|HIE1@[EE1HmHEt$H$P0t$$iHH|HI;[EE1(HT$ LH5.&LHdZ.L\^IZ=E11E1Yg[FJL&^H@f.AWAVAUATUSHHHH x"dH%(HD$81H<$Hz&HH=z&HL$0HD$ H|$(LFI8~.II HF(HD$H^ Hn"MHH"HHD$Hm[HD$_[D$ZH YH& L(hE111AHHAHIjH;t"HG& H(hE111HAHHHH;-"H-|&H=&HXHIHIVH5f&HHjLHHI.L-L|&H=&LuXHIYHIVH5|&HHLIMI.fWWHIbXHI=H1Lp HXHL [HI I/uImZXHILp1HHZHIHmLIm1L;5"L;5"eIHPHIH$HT$IH5d"LHh HEIHx\HIHmI,$ZHtH+u HCHP0LDMM@H=!11fcYH zH=r k1HL$8dH3 %(7HH[]A\A]A^A_ffWf.D$H$HT$H5"L$D$Hh HEHHxHIHm3HEHP0$fL;59"LWZ nfH "HD$H"HHD$@ID$LP0IFLP0IFLP0BIELP0IGLP0|IELP0HEHP0IFLP0HEHP0 LrHIE1Z E1HmHEt$H$P0t$$iHrHIY HHHD$0H[IW HB@HH$H5_k&HxF'I޾pW 1W E1W E1W W E1W W E1HB@HH$4}W E1nW H5j&Hx FMcW 1E1E11E1GW E1(HT$ LgH5!&LHQGV.LAI,W E11E1YW JL@H@f.AWAVAUATUHSHHHdH%(HD$81H%"HH<$HD$ HD$(HD$0LFIcIHF(HD$Hk H[H`>HD$R>D$=H6yHIuI.I/rx;HIpLh1HH=HIHmdI/nL;-۫"L;-"IEHPHIUnH$HT$IH5+"LLh IEMHxSHIImI,$QHtH+u HCHP0L^H5^&HI9HHD$ LCfH=Rc1I[RH ]H=c N1HL$8dH3 %(YHH[]A\A]A^A_fDH"HD$fWf.D$?H$HT$H5"L$D$Hh HEHHx跇HIHmHEHP0L;- "eL:VSH fID$LP0H@HP0H@LP0MIELP0IELP0+IGLP0IFLP0eIELP0HEHP0I/IGLP0DIELP0H5RU&H=sd&1:HHHPRH+R? E11E1@E1E1Mt ImMt I. Mt I/H [H=`E1LMQWHUHIrRH LTHIE1RH E1HmWHET$H4$P0T$4$:@H51T&H=Zc&19HIH7QImti'SI E1IET$L4$P0T$4$IFT$L4$P0T$4$IGT$L4$P0T$4$IELE1P0'SI {HCHP0V74LfIt(ItIHF(HD$0HC HD$(HCHD$ Hd1IIIMMfHD$0H\$ Hl$(HD$E1SH tSH e SH VHB@HH$SH E1.RH {E1RE 1gRF E1USH E1H5Y&H4HHD$(IMH5:W&H4HtpHD$0IHB@HH$RH E1s>SJ MME1E1便R@ E119HT$ L]H5B&LH=tJRrH=]1ACARJR? E1E1(#SI Lq7IoLa7HfAWAVAUATUHSHHHdH%(HD$81H"HH<$HD$ HD$(HD$0[ LFIIHF(HD$Hk H[H4HD$4D$'4H2H`& L(hE111AHHAHI H8H_& H(hE111HAHHH H8H!V&H=b_&HJ2HINHIVH5X&HH LHH I.TL=U&H=_&L1HIHIVH5cV&HH LIM I.fWo1HI I2HI I$1Lp L`HL4HI ImI/1HI Lp1HH;4HI H+I/L;5\"L;5""?IHPHIYHT&H=]&H0HIgHIVH5@W&HH LIM I.-H&T&H=g]&HO0HIHIVH5T&HH+LHHrI.fW/HIe0HIaHE1Lp HhHH2HI\H+ImY0HILp1HL2HII/ImkL;5"L;5"IHPHIYH$HT$IH5n"LHX HIHx7HI3H+!I,$nHtHmu HEHP0LbH5T&HI.HHD$ ULCfDH=;X1>9OH RH=ea C1HL$8dH3 %(ZHH[]A\A]A^A_fDHi"HD$fWf.D$Wf.D$H$HT$H5L"L$D$HX HHHx|HI1H+HCHP0L;5ٞ"Lk/P; E1E1OID$LP0L;5"L/LP= E1E1Mt I.iMt ImyMI/IGT$L4$P0T$4$iH@HP0aH@LP0 IFLP0IFLP0IGLP0?IELP0%IFLP0HCHP0I/QIGLP0BDIFLP0fIFLP0IELP0^HCHP0CIELP0IGLP0kIFLP0HCHP0H5JI&H=;X&1.HHHFHmtO1 1E1fDH OH=Z^E1@Mhn@LIHINE1O; E1H+HCT$H߉4$P0T$4$DHHHIO; eDH5QH&H=RW&1-HIH/EIEP< HHIEIELP0P< H(HHI 0P= <DHHHI+P= DH5G&H=V&1-HI5HwDIE[P> HHIE`IELP0[P> GIFT$L4$P0T$4$zIET$L4$P0T$4$jH5F&H=U&1v,HH4HCHmtO3 1E1HEHE11P0O3 HEHE11P0O1 'LfIt(6ItI/HF(HD$0HC HD$(HCHD$ HC$IItUItpMM-HD$0H\$ Hl$(HD$QO; E1E14HP= 4H5L&H6(HHD$(6IM~H5J&H(HHD$0InHB@H"H$O; O; E1O; E1CP= 2P= E1oE15P= a7P= R-P= E1 CLc)I#f.AWAVAUATUHSHHHdH%(HD$81H"HH<$HD$ HD$(HD$0[ LFIIHF(HD$Hk H[H&HD$&D$'&H$HR& L(hE111AHHAHI H8HQ& H(hE111HAHHH H8H!H&H=bQ&HJ$HINHIVH5J&HH LHH I.TL=G&H=Q&L#HIHIVH5cH&HH LIM I.fWo#HI I$HI I$1Lp L`HL&HI ImI/#HI Lp1HH;&HI H+I/L;5\"L;5""?IHPHIYHF&H=O&H"HIgHIVH5@I&HH LIM I.-H&F&H=gO&HO"HIHIVH5F&HH+LHHrI.fW!HIe"HIaHE1Lp HhHH$HI\H+ImY"HILp1HL$HII/ImkL;5"L;5"IHPHIYH$HT$IH5"LHX HIHx7HI3H+!I,$nHtHmu HEHP0LbH5F&HI HHD$ ULCfDH=cJ10wJH DH=,J51HL$8dH3 %(ZHH[]A\A]A^A_fDHi"HD$fWf.D$Wf.D$H$HT$H5"L$D$HX HHHxnHI1H+HCHP0L;5ِ"Lk!NK) E1E1OID$LP0L;5"L!K+ E1E1Mt I.iMt ImyMI/IGT$L4$P0T$4$iH@HP0aH@LP0 IFLP0IFLP0IGLP0?IELP0%IFLP0HCHP0I/QIGLP0BDIFLP0fIFLP0IELP0^HCHP0CIELP0IGLP0kIFLP0HCHP0H5;&H=;J&1 HHH8HmtJ 1E1fDH AH=!GE12Mhn@L;HINE12K) E1H+HCT$H߉4$P0T$4$DH:HI-K) eDH5:&H=RI&1HIH/7IE]K* HHIEIELP0]K* H(:HI nK+ <DH:HIiK+ DH59&H=H&1HI5Hw6IEK, HHIE`IELP0K, GIFT$L4$P0T$4$zIET$L4$P0T$4$jH5\9&H=G&1vHH4H5HmtJ! 1E1HEHE11P0J! HEHE11P0J LfIt(6ItI/HF(HD$0HC HD$(HCHD$ HCIItUItpMM-HD$0H\$ Hl$(HD$Q/K) E1E14K+ 4H5&@&H6HHD$(6IM~H5<&HHHD$0InHB@H"H$9K) 7K) E14K) E1K+ pK+ E1oE1sK+ auK+ RkK+ E1<}K+ 1AK) "EK) E1HB@HH$E1JK) HB@HH$+K' 1uK( fHB@HH$8E1I޾K- E1MME1E1J" E11YL-HNLIHT$ L#BH5%LH"fJYK* J! E1J E1H=A1A']JWLHK, CLcI#f.AWAVAUATUHSHHHdH%(HD$81H<$H"HH=8&HD$ HD$0H|$(D LFIItjIH=A1'HH ?;H=@h+,1HL$8dH3 %( HH[]A\A]A^A_fDHF(HD$Hk H[HwHD$iD$HmHC& L(hE111AHHAHI H8;HTC& H(hE111HAHHH H8H9&H=B&HHIHIVH5v<&HH? LHH I.L=\9&H=B&LHIHIVH59&HH LIM I._fWHI] HI I$1Lp L`HLHI Im#I/ HIt Lp1HHHIi H+I/L;5"L;5"IHPHIHH8&H=ZA&HBHIHIVH5:&HHtLIM> I.|H7&H=@&HHI*HIVH5[8&HH.LHHI.0fWgHI*AHIHE1Lp HhHH{HIH+ImHILp1HL3HI I/ImL;5S"L;5"IHPHIXH$HT$IH5"LHX HIHxΏHIH+xI,$HtHmu HEHP0LfH"HHD$@Hi"HD$hfWf.D$f.D$9H$HT$H5\"L$D$HX HHHx`HITH+_HCHP0PL;5ق"LkIE1E1@ID$LP0L;5"mL^IE1E1@H@HP0fH@LP0IFLP0FIFLP0IGLP0IELP0IGLP0HCHP0IFLP0!IFLP0uIFLP0IELP0HCHP0IGLP0Im9IELP0*@IFLP0IfHCHP0yH-HI IfH 4H=9E1$M9?@Lp-HIE1IE1H+uHCt$H߉$P0t$$Mt I.wMt ImGMnI/dIGt$L$P0t$$GH5-&H=;&1HIHw)IIHHIIFLP0IHx,HIIFDHP,HIBIDH5a,&H=:&1kHIH(IJHHIJIFLP0J1IEt$L$P0t$$IFt$L$P0t$$lH5+&H=M:&1HH_H*(HmtXHI1E1H5+&H= :&1HH-H'Hmt2.I1E1kHEHE11P0HIMHEHE11P0.I/* LfIt(mItIfHF(HD$0HC HD$(HCHD$ HWIIIMMHD$0H\$ Hl$(HD$IIyI1vHB@HyH$IE1M~H5.&H HtHD$(IMeH5t.&H HHD$0I6IE17E1I޾JE1=HB@HH$vHB@HH$HB@HjH$E1IIE1E1IIII.IIE1E1tIE1bIE1PI#H5H-&HI HHD$ hLC]MMLE1E1_IE11IsDIE1a*IE1OI@L IL HBHT$ L3H5%LH*&H=3&HgHIHIWH5*&HH LHHI/HIHEHh1H'&LHHIG  HIdH+pI/z|HILp1HLHIxImPI/ZL;5v"L;5u"IHPHI<H$HT$IH5Yv"LLp IMHxZHII.I,$iHtHmu HEHP0LeH5+&HIHHD$ GLCfH=.1DH (H=z.{1HL$8dH3 %(2HH[]A\A]A^A_fDHu"HD$fWf.D$7f.D$iH$HT$H54u"L$D$HX HHHxRHIH+HCHP0L;=t"LxETDID$LP0L;5t"LKEE1@Mt I.qMI/IGT$L4$P0T$4$iH@HP0H@LP0=IGLP0IELP0IELP0fIFLP0KIELP0HCHP0sIGLP0IGLP0IGLP0@HCHP0I/IGLP0wDIELP0I/IGLP0DIFLP0IFLP0H5Z &H={.&1HHHXHmlD1E1fDH %H=+E1M@LHHI~E1^EE1H+uHCT$H߉4$P0T$4$MImIET$L4$P0T$4$DHHIYE=DH59&H=j-&1HI&HGIEHHIIGLP0EHHHIEDH HIEE1fH5&H=,&1;HIHIEHHIBIFLP0E)IFT$L4$P0T$4$rH5&H==,&1HHHHmt E1E1HEHE11P0 EHEHE11P0D^LfIt(0ItI)HF(HD$0HC HD$(HCHD$ HIItlIMM[HD$0H\$ Hl$(HD$JEHB@HVH$EE18H5$&HgHHD$(.IM~H5 &HAHHD$0I[EE;E1SE1cEkE~`EE1lM[EE1E1rHB@HH$koEE1,tEE1E EE1HB@HH$EHB@HH$E1EE1E1!EE11MM-EE1#DE1LqHLaI ELBHFEHT$ Lg&H5%LH(DLI`H=,&1A DpAWAVAUATUSHHHH (m"dH%(HD$81H<$H&HH=&HL$0HD$ H|$(LFI0~.I IHF(HD$H^ Hn"MHHl"HHD$HHD$D$dH6H?(& L(hE111AHHAHIH8H'& H(hE111HAHHHH8]H-^&H='&HHI+HIUH5!&HH'LHHIm L=&H=B'&L*HI~HIUH5&HH+LIMImHIHHX1HT&LLHIE HII.Im>HIYLx1HHHI*HmzIm_L;=j"L;=fi"cIHPHIMH$HT$IH5Zi"LLx IMHxvHII/I,$bHtH+u HCHP0LDMM@H="11eCH H=u" 1HL$8dH3 %(mHH[]A\A]A^A_ffWf.D$H$HT$H5xh"L$D$Hh HEHHxFHIHm3HEHP0$fL;=h"L DvE1Hh"HD$Hh"HHD$@ID$LP0H@HP0H@LP0@IELP0IELP0IELP0iIFLP0NIELP0HEHP0wIGLP0IGLP0LhHIrDvE1E1HmHET$H4$P0T$4$qHHIDvE1DDH5&H="&13HIkHI/tY/DwE1E1E1MtImtVMtI.teMtI/twH H=E1 MIGLE1E1P0/DwIET$L4$P0T$4$IFT$L4$P0T$4$~IGT$L4$P0T$4$lH5&H=!&1;HHHH+tCnE1E11E1HCHP0LfIt(&ItIHF(HD$0HF HD$(HFHD$ H3IHIMHHD$0Hl$ H\$(HD$E1CtE11QFDxBDvE1DvE1Dv| DvE1jDvE1XHB@H=H$DvE1H9H5$&HHtHD$ HHH5>&HnHtHD$(HHH5&HHHtwHD$0HHB@HtPH$CuE1E1%E1CoE1E11xCnE1E1E1LHxHT$ LH59%LH HY"HD$XHD$(HHH`HphH@pHHL$Ht$HD$tHHD$HtHHD$HtHHHL$L5:Z"tIGHD$XHT$PL9%HHhH}#HA(Hp#H$HLЅHT$PH*#IHHD$PLIuH HIGL9H;!X"HPhH+#HBH#H#HLHHD$X IGL9H;W"H@hH"H@H"HLHHD$PIGHT$XL9H@hH"H@(H"HLЅHT$XH*H|$XHGP0@H|$PHGP0HIWH H9$"H"IWHHHD$XIGL9 I;_"IGJ HHD$PIGHT$XL9-I;_!IGJ< HIGJ H//HGP0#HH*I9GO!HF!IOHHI'HT$PH*HV"H MU"HD$PI9HD$@HL$HI9IHPHIIWH5M&HH(LIM(L;t$@L;t$HIHPHIFFIWH5s &HHq"LIMI"IVH5/ &HH"LHHD$P!I.J H|$PHGH$HGH#HD$H/ IWHD$PH5&HH $LHHD$P"HPH;iU"sH;S"HRhH!HRH!1HIM1HT$PH* IFHD$PH"IFHD$H|$"I.l IWH5&HHp#LIM"IVH5&HH.#LHHD$P"I. H|$PHGHP"HoH HT$PH* L5)&H=j&HD$PLIHHHD$PHHH5&HHy HIMQ HT$PH* HHD$PHHD$P!HHD$XHT$PHD$PHPHHD$P8 L%a&H=&LHIHIWH5g&HH* LHHD$` I/ HT$`H5&H|$P<HT$`H* HT$PHt$XLHD$`HHD$`I. HT$XH* HT$PHD$XH* HD$`HD$PHD$`H5&HPHH$HHHHD$`HHH5_&HHHHHD$PHT$`H*e H|$PHD$`HGHQLwMH/ I} HD$PH5.&HWHHHD$0H|$0I} H5&HWHH1HHD$PH5 &1HHHD$`HT$PH* HT$`HD$PH* H#P"HD$`HHH`HphH@pHHL$8Ht$ HD$(tHHD$ HtHHD$(HtHHH H~uL|$MLLHL|$HD$It$HHD$HL$HLL,LfLHLXLHLJL|$HuHO"HH\$`HH|$8HD$`tHL$8HHD$HHHy HL$ HtHHD$HHH/ HL$(HtHHD$HHH% Lt$0H5%1LHIHD$HHIT HH;l$@H;l$HDHEHPHHU EHH$HtHHQHL$HHu HH@P0HL$hdH3 %(HHx[]A\A]A^A_H|$PHGP0IFLP0fHD.fDHDfDL;5 N"LsE15H|$XHGP0H|$PHGP0H0HHD$XRtH$ IWH5%&HHjLHHD$P)H5%HHHD$XHT$PH*H:M"H|$XH K"HD$PH9HD$@H9HL$HŃHHPHHHD$XIWH5%HH.LHHD$XH;D$@H;D$HHHQHHHD$XH-%H= &HHHHD$XHHH5&HHHHHD$PHT$XH* IGL5qL"HD$XL9gH;J"H@hHEH@H81LHHD$XHHD$`HT$XH|$PHHD$XHP12HHD$XHT$PH*% HT$`HD$PH* I} HD$XHD$`HD$XH5&HWH$HHHD$H|$I} H5&HWHH~HHD$XH5Z&1HpHHD$`WHT$XH* HT$`HD$XH*c HI"HD$`HD$(HHH`HphH@pHHL$ Ht$HD$tHHD$HtHHD$HtHHHF L$IGHD$`L9 H;H"n H@hHAH@H4HLHHHT$`IGL9o HHhHHA(HHHLЅHT$`H*IGHD$`L9` H@hHH@(HH$HLЅIHB IuHHIGL9^H;G"9 HPhH1HBH$HHLHHHT$`- H5vH"H<$HT$`H*}H|$`HGP0lDH|$P$HWR0$mfH|$XHGP0H|$PHGP0IFLP0H|$PHGP0]IFLP0fIFLP0H|$PHGP0H|$PHGP0KIGLP0IFLP0HT$XH*DH|$XHGP03fDH|$`HGP0H|$PHGP0H|$`HGP0H|$PHGP0H|$`HGP0EH|$PHGP0HCHP0 HL|$Ld$LLLHIHMHuHdHD$LHIIHIIMHuH9\HAPHAHP0fHAHP0HAHP0xHhHIxsH$UDHEHP0H|$0HGP0IFLP0afH|$PHGP0HxAH@HHIf.HxHHHIDH;= E"$H|$Xf.H|$t#H\$HHD$ HHHu HCHP0H\$HtHHD$HHHu HCHP0H\$HtHHD$HHHu HCHP0L4$H5J%1L0HIHD$HHI\HH;MD"H;C"t4HHPHH> H$HC"1fH;C"tHoHD$XfuHD$`HtHHQHH^HD$XHD$`HtHHQHH HD$PHD$XHtHHQHHH NH=FHD$P/HT$`Ht$XH|$P+ HIHD$PL<$1LHLHD$PIFHD$XHHD$XIF HD$`HHD$`IF(HIHD$ HHIHcH;B"H;yA"HHPHH  I.hHT$PH*AHT$XHD$PH*HT$`HD$XH*HT$Ht$H|$HD$`lH;A"]HkQ@H<$HGP0HCHP0AufDLHHD$PsH$E1HD$PHtHHQHH@HD$XHtHHQHHMt I/HD$`HtHHQHHH H=މ1,H<$HGP0HB@HH$H<$HB@HIH$fH|$`HGP0;HIWHyHDH9 H IWHHHT$`I;_ IGJ HIGHT$`L9HHI9Gf H] IOHH}HWHHUr H8HWHHP H""HHx`HphH@pHH|$HHt$8HD$@tHHD$8HtHHD$@HtH致HD$`IF1HLl$hMIHl$PH\$PHhfH(H0I8H@(I@H@I@HXH(H0I@H@(II9aI@HH0I8H0H0KI8H@I8HT8H(HI8HcI H0FL;%!"ULWFafDIFLP0PH@LP0HH@P05IELP0IFLP0nIELP0I.ZIFLP0HCHP0I.|ۀ8CH(IHI8HcI H0I9MH\$PLl$hH|$`虫H|$HtH|$HHHD$PHHH H|$8HtHHD$HHHH H|$@HtHHD$8HHH Hl$0H5%1H*IHEHD$8HHHE<MvL;d$ L;d$(I$HPHI$IELHIEHHHr MtI.u IFLP0MH+H\$HHD$ HHHMtI/u IGLP0HĈL[]A\A]A^A_fD2HcHL`1HH}HIH+ I. L;%!L;%d!P I$HPHI$ H-®%H=%HHHHHSH5%HHHIMH+ H-f%H=%H菊HIHIT$H5ï%HH LHHUI,$l HIHD$1LHHID$H%HID$ &HIlH+ I,$ 蛊HI0!Lx1HLތHI$!I. I,$ L;=!L;=!G IHPHI HD$8Lx IH@HD$@HD$0H;! H%H==%H%HIHIVH5®%HHxLIMI. 蕉HHHD$01HLHHEH!HHE HH I,$> Hm# H; L%1HkLHT$ H޿AHIUH8 MfHO%sH{ I9X IWH5V%HHLHD$H|$IWH5-%HHLIMuH5%1LˊHI<$HWHI$^H8HWHH<HS!HHx`HphH@pHH|$8Ht$(HD$0tHHD$(HtHHD$0HtHHD$HIFE1HLl$PH\$XMHIHl$@c@H(H0I8H@(I@H@I@HH(H0I@H@(II9I@HH0I8H0  KI8H@I8HS8>H(HI8HcI H0EMMH=1bH /H= 1H\$xdH3%(VHĈ[]A\A]A^A_8;H(IHI8HcI H0I9Ll$PH\$XH|$H|H|$8t H|$8HHD$@HHHuHGP0H|$(HtHHD$8HHHuHGP0H|$0HtHHD$(HHHuHGP0Ld$H5%1LHI$HD$(HHI$ HH;-!H;-!S DHEHPHHU5EHIHHHHHMtI.u IFLP0MI/H\$ HHD$HHH>MtImu IELP0L'HcHHWHI H8HWHH H!HHx`HphH@pHH|$0Ht$8HD$@tHHD$8HtHHD$@HtH菃E1HHD$H~$fL$ LD$(KDII9uH|$Ht|H|$0tH|$0HHD$ HHH4H|$8HtHHD$ HHHH|$@HtHHD$ HHH Lt$H5%1LHIHD$ HHI HhH;-"!H;-! DHEHPHHU EHIHHHHHu HCHP0MI/EIGLP06HPHR0IT$HD$(LR0HD$(L;5N!OL@c T HCHP0L;% !L蜀cE1 HH0H;0HHH0I@H0H0HH0H;01HHH0I8H0H0\L;=s!L dE1E11B @HCHP0yHCHP0H@LP0HH@P0MHCHP0ID$LP0HCHP0IGLP0IFLP0IFLP0*IFLP0ID$LP0AHCHP0IFLP0IGLP0IFLP0(HCHP0VID$LP0HCHP0ID$LP0ID$LP0IFLP0IGLP00IGLP0PHCHP0H@LP01HCHP0HEHP0ID$LP0IFLP0LH@0I8H@(I8H(H+0H0H@0I@H@(I@H(H+0H0H5j%H=K%1~HHH(HE*HHHEIH H=5裐HHE1H;-!H}AH K%1LHt$ HI H8H%Av(E1E1IV0HxHD$D$$HHH8GIWHkH5%HHLHD$H|$*IWH5%HH LIM H5F%1L\}HH. I,$H*H!HHx`HphH@pHH|$(Ht$0HD$8tHHD$0HtHHD$8HtH|HD$HIFE1HH\$XH\$@Ll$PMMII8HH0I0H0  xJDAEE1IE +xDH(H0H0H@(AE;U}OIcItH0H@H0Pt8EH(AHR8HcR H0E;U|IM98MH\$XLl$PH|$HtH|$(t H|$(HHD$@HHHuHGP0H|$0HtHHD$(HHHuHGP0H|$8HtHHD$(HHHuHGP0Ld$H5%1LQ{HI$HD$(HHI$HH;-l!AH;-1!DEHmEUMfDH|$HGP0HEt$(HT$P0T$t$(HBHP0fHD$(ID$LP0HT$(BH@HP0H@LP0IH|$HGP0$H5%H=£%1CzHI H蟑Im bE1HD$ fDH $H=E1H|$ qHcHHH$HHH5%HHHIMH$H*(HDŽ$aHH$ːH$Lh1HdHI5H$H*C(H$HDŽ$H* (IUHDŽ$H5N%HHLHH$Im'+aHIH$H$1LHDŽ$IEUcHH$rH$H*'HDŽ$Im'H-%H$H=%HDŽ$HHD$P_H>HH$HHH5%HHHIMH$H*Y'LHHDŽ$]IMHQHIU '\H-%H=>%H&_HIBIHIUH5%HH9LHH$9ImbHSH5%HHFHIMԈH-o%H=%H^HHHH$HxH5%HHLHHH$HH$H*HDŽ$^HH$H$H$HLhHDŽ$HP 1 aHH$H$H* H$HDŽ$H*H$H;=!HDŽ$H;=!ŃHHPHHb)HDŽ$nH-%H=P%H8]HHXHH$HWH5%HH HH$zH$H*OH-%H=%HDŽ$H\HHWHH$HWH5#%HHHH$!yH$H*NHSHDŽ$H5%HHrHHH$;q\IH$M*H$1IELHDŽ$ _HH$ԈH$H*6HDŽ$Im6H$H5H%HWHHIMH$H*L6HDŽ$\HH$H$Lh1HO^HI2H$H**6H$HDŽ$H*5HD$PHDŽ$LHHH$XHHH$LH;!H;=!95HHPHHiMHDŽ$AIEL$H$H*LHDŽ$ImLH$HEH$H*(MHL$PHDŽ$HHD$HHHLHl$PfH% H(hE1AAHHHI|H$H8!H+!H$H;!HDŽ$|HSHkH5L}%HH\}HHH$\}H5z%HWHI}H$H* !L;-]!HDŽ$L;-!t DIEHPHIU !ER8HSH5/{%HHHIMHt$@LtVHH$^Im0!H$H;=!H;=w!| AHAHPHH EHDŽ$8L%{%H=%LWH9HH$HHH5O|%HHHIMH$H* IUHDŽ$H5z%HHLHH$tImt H5x%1HCUHIpWHH$H$Lh1H&ZHIH$H*H$HDŽ$H* L;-+!HDŽ$L;-!zDIEHPHIUE8Ht$hHFH~xHFHxH]>HUHDfW ff( HH9\f(Xf(\\u\ fTUHIt}Ht$PHSHH${Im H$H;= !H;=!HHPHHHDŽ$vJH$HH;! H-y%H=Q%H9UH:HH$HHH5x%HHvHIMH$H*HDŽ$UHH$uH4$HH$HpVHH$uH-Vx%H=%HTHV9HH$HxH5Gy%HH$vHHHH$H*H5y%H$HHDŽ$WIHmH$H$LVHHD$(NImH$H*H$HDŽ$H*H $HDŽ$HHD$HHHL;5!L;5!Q&H;!)#HSH56y%HH yHHHH5%1HVHH$HmL$HDŽ$ID$H;!z=H; !mEHhhHdH}dHEHLHHpLUHdH!z@IWH5=v%HHyLHH$yH5s%H PHH$H$H*@H$H;=A!HDŽ$H;=!8HHPHHWHDŽ$7IWH5)t%HHKLHH$ HPH;(!+H;!1HRhHW]HRHJ]1HHH$sMH$H*iH$H;r%HDŽ$HDŽ$HD$@H5o%H=Q}%1SHH$H)kH$H*4QH4$H$Aj:HDŽ$HD$HAu1E1HD$8Ht$(HD$XE1HD$0HD$E1HD$HD$`HD$HD$ H$HD$PHD$hHtH0HVHH+Mt Im+H$HtH0HVHH+H$HtHHQHH,Ht Hmx+H sH=;zDDE1d!IHo!HUH;=Y!PP[H$HHPHHH$HGP0vDH$HGP0@IELP0H$H*EH$HGP01H$HGP0@IGLP06fH$HGP0@H$HGP0}@HD$PHD$hfL;59!LOH$HD$HA"<A1E1HD$8HD$XE1HD$0HD$E1HD$HD$`HD$HD$ LEf.H=Mx1X]B9H qH=xkb1H$dH3%(`NH[]A\A]A^A_H9!@H;=)!iNJyH$RH$HGP0@H$HGP0@H$HGP0<@IELP0fL;-!L+NApH $H$AQ;HD$HA1HD$8HD$XE1HL$(HD$0E1HD$HD$HD$`HD$HD$ H$DH;=!wMAeH$HHPHHfH$HGP0R@Hl%H $HHD$(H1HVHt$HHHAHP0Hl%HD$(Ht$@H|$( IHHdH;a!H;-'!<DHEHPHHU++ErFH;!3;L%vn%H=w%LJHHAHHUH5p%HH\HIMсHmx1H5k%HDHHHViJHH$hHh1HL,MHHIm1H$H*1Ht$(1HHDŽ$GHH$(cHm0H$H;=!H;=˹!(ŃHHPHHh0HDŽ$@Hj%HHD$HHSH5io%HH3HHH$H5v%1H*LHHD$H$H*0HDŽ$H+[0Hl%H=u%HHHHAHHUH5Kj%HH#iHHH$hHm/(IHHiH$HHEkJHIHHl%H=Au%H)HHAHH$HxH5!m%HHHHH$&H$H*{/H$H5m%LHDŽ$JEH$H*/H$LHHDŽ$JHH$H$H*/HDŽ$Hm/Im/H$H5Aj%HDŽ$HPHHD$`HHHH$ܓH5s%1HIHHD$^H$H*./L|$pL|$HE1HDŽ$HD$0E1HD$HD$ fDHt$(1LiDHIPH;!L;-!IEHPHIUF9H|$XH5mi%HWHHLIMUH|$(LIHH$TFHH3WH$1HLHDŽ$HEHHH$VImHmHL$HtHHD$8HHH9H$H5f%LHDŽ$HD$$CHHH$TH;q!H;=7! HHPHH-HDŽ$HD$HPHjhHTHE HSL;=!J LBHHH%H~@H+HH%H%HE 1H|$HHH$He%H|$|??H$H* HDŽ$Hg%H=q%HCHHH$HHH5i%HHTHHHATH$H*mHDŽ$GDHH$cTHD$1HHH$HFwFHHIHmH$H*HL$ HDŽ$HtHHD$8HHHHCH;ʹ! H;@!HhhHCH}CHEH΁HH%HpHUIMLd$XL|$H1H\$ L|$pA=H\$H$AHD$8HD$E1pL;-!LDLd$XL|$HAf=H\$L|$pAH$1HD$8HD$E1 DH;=!!CJH$HHPHHH$HGP0DIELP0`fHEHP0IELP0HEHP0H$H*H$HGP0H$HGP0@HAHP0fHAHP0HCH;ò!HSHBH9AHAHSH HI@LH>HH$JIm H+ H$H58c%HDŽ$HPHHD$ HH HH$I@HIIHD$HIE5BHH&IHb%H5b%HC:3H$HLBHH$FH$H*E HDŽ$Im Hm Mt I. HEc%H$H=~l%HDŽ$HHD$8U?H HH$HHH5a%HHFHHHBFH$H*R HDŽ$?HH$PH\$8HH$HX@HIkPH!H5a%LHHH$A2H$H*f H$LHHDŽ$wAHH$DHm H$H* HDŽ$Im H$HPH;W!HHxHlHXH@ H$HH$HH$H* HDŽ$HL$0HtHHD$HHHH Mt I,$ L$H5_%HDŽ$IT$HHHLHH$GH5.j%1HD@HH$EH$H* H$HDŽ$H*c H|$8HDŽ$H5^%HWHHXHH$DZ=HH$cDI$1H$H$Lf?HILH$H* H$HDŽ$H*} Ht$8HDŽ$HHD$0HHH< IVH5^%HHLLIMOLL8HH$MCIma HD$H@HhhHs<H}0h<L;=ɬ!L+?HB HHD$0e7HT$0IH* I H$H;=z!>H HHD$07HL$0HH)G H HH? LH? LLH|$U0pH$H*wIVHDŽ$H5(]%HHx<LHH$<HL6HI;H$H*)HDŽ$I/+MH\$0fDL;-a!yL;AjH4$H$A;HD$HA1HD$8HD$XE1Ht$(HD$0E1HD$HD$HD$`HD$HD$ H$DH;=!N;GyH$HHPHHH$HGP0DH;=i!:_xH$HHPHH H$HGP0DH$HGP0@IELP0DfH$HGP0@H$HGP0@H$HGP06@H$HGP0@H$HGP0*@IUD$LR0D$H$HGP0@IELP0fH$HGP0@HCHP0EfH@LP0#IELP0H;=٨!n9tzH$HHPHHH$HGP0DHH$HGP0@IELP0fH$HGP0`@H$HGP0@IELP0}IELP08H$HGP0@H$HGP0@H$HGP0@H$HGP01@H$HGP0@H$HGP0@H$HGP06@H$HGP0@HEHP0H$HGP0G@H$HGP0@IELP0HAHP04IELP0H$HGP0@ID$LP0HEHP0IELP0HEHP0QHRHH$DE1oH$HGP0@HEHP0fIELP0H$HGP0@H[fH$HGP0@HEHP0fH$HGP0@IELP0fH$HGP0@HFHP0fH$HGP0o@H$HGP09@H$HGP0@H$HGP0S@IELP0H$HGP0u@H$HGP0@IGLP0IFLP0mHAHP0VID$LP0UHBHP0nHD$0HAHP0HT$0HhOHH$GLd$XLd$8A=L|$HH\$AL|$pH$1ME1HD$8HD$E1X[2HIfD;2HHHEHHT$0H|$HHT$0*uHuMLHIIfDH;-!Hs2AH$HD$HA<AE1HD$8HD$XHD$0E1HD$HD$E1HD$`HD$HD$ 5DH|$XH5TS%HWHHeHH$e0HH$HHSQ%Ht$@HH$HBHH$Hp 1HIXGH$H5R%H2,H$H$Lr2HHD$FH$H* H$HDŽ$H* HDŽ$ImS HD$HHD$XE1HD$0HD$E1HD$HD$`HD$ H$H;!IWH5R%HHtLHH$tH5O%HZ,HH$sH$H*'H$H;=!HDŽ$H;=J!HHPHHHDŽ$;HD$HD$8HIHL$@HtHHD$pHHHHL$hHtHHD$@HHHHL$PHtHHD$@HHHH $HtHHD$@HHHHL$ HtHH$HHHMt I,$HL$HtHH$HHHHL$HHtHH$HHHHL$`HtHH$HHHHL$HtHH$HHHzHL$HtHH$HHHiMt I.jHL$0HtHH$HHHYHL$XHtHH$HHHHHt$8HtHH$HHH7I/tAHL$(HtHH$HHH HtH+u HCHP0LIGLP0@HAHP0HAHP0HAHP0HAHP0"HAHP03ID$LP02HAHP0DHAHP0UHAHP0fHAHP0wHAHP0IFLP0HAHP0HAHP0HFHP0HAHP0H$H;՛!Ht$L2)HI)HD$8H;=!L.,pH$HHPHH:H$HGP0&DH$HGP0H$HD$HE1HD$XHD$0E1HD$HD$HD$`HD$ H-qM%H=V%H)H#HH$HHH5zM%HHBHHH$gH$H*iH$H|$HDŽ$'GH$H1HVHH HDŽ$MH|$H5M%HWHHGHH$FH5JG%1H+HH$TH$H*Ht$HDŽ$HHD$8HHHtPH$HDŽ$HD$fHCHPH9(H(HLHI]DHFHP0@H$HGP0g@H|$H5K%HWHHSHH$SH5FI%H%HH$HSH$H*H$H;=ט!HDŽ$H;=!{ŃHHPHHHDŽ$H-J%H=$T%H 'H!HH$HHH5L%HHeHHH$dH$H*HDŽ$(HH$ eIWH5OL%HHPeLIMeH5'L%H$L_)$ImH$H5D%H$@)HHD$8BH$H*cH$HDŽ$H*1Ht$LHDŽ$y$HI$H5R%H|$8H I$ImHD$8HIDHx1H@HHH$RH;=!+'^SH$H$HGP0-H$HGP0IELP0fLH$HGP0!H$HGP0H;= !x&jH$aIELP0H$HGP0 H$HGP0_H$HGP0-HB@HcH$H|$8HB@H!?H$H|$ fH$DT$|DL$pHGP0DT$|DL$p0HEDT$|HDL$pP0DT$|DL$peDIEDT$|LDL$pP0DT$|DL$pDH$DT$|DL$pHGP0DT$|DL$pH$DT$|DL$pHGP0DT$|DL$pHAHH$H4$H$A:HD$HAz1E1HD$8HD$XHt$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PHX@HH$H $H$A:HD$HAz1E1HD$8HD$XHL$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PH?HH$ٿH $H$A:HD$HAz1E1HD$8HD$XHL$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PXH>HH$`H $H$A:HD$HA{1E1HD$8HD$XHL$(HD$0E1HD$HD$E1HD$`HD$HD$ H$H;=ɑ!^"oH$H5!?%H=L%1k#HIVH:ImH4$H$A`;HD$HA1E1HD$8HD$XHt$(HD$0E1HD$HD$E1HD$`HD$HD$ H$fDHEHP0Hx+H@HH5)>%H=K%1{"HH$>H9H$H*qH $HDŽ$A;H$HD$HA1E1HD$8HL$(HD$XE1HD$0HD$E1HD$HD$`HD$HD$ H$fDHeHHEHHT$8H|$HHT$8pYH¹HHLH\$A1E1HD$8HD$E1H$D$8HWR0D$8H$HGP0H$HGP0H$HGP0H$HGP0IELP0>HB4HH$޷H $H$A:HD$HA|1HD$8HD$XE1HL$(HD$0E1HD$HD$HD$`HD$HD$ H$fDH3HIH4$H$1HD$HHD$8E1HD$0HD$A:Ht$(HD$`A|HD$HD$ E1H$HD$XHD$aIT$HBH9{'Hr'IT$H4HH$L(HHD$ `H$H*HDŽ$I,$aH|$XH57%HWHHj^HH]YHH$m]H$1HHH$HFHH$\HmH$H*H|$ L$HDŽ$HDŽ$H56%HWHH_HH$^HH$^I$H$L`HH^H6%H566%HH$H$HHHD$R]H$H*nH$HDŽ$H*8HDŽ$HmH-6%H=#@%H HIHIUH59%HH\LHHDZImzHIZHD$HIEHH$[ZH5!HHH$HH58%vN H$H*H$LHHDŽ$MHH$,^HmqImNH$H*$HL$HDŽ$HHD$HHH;H$HD$HE1HDŽ$HD$XHD$0HD$HD$HD$HD$`@ID$LP0H$HGP0d@H$HGP0@HEHP0IELP0MHEHP0H$HGP0@H$HGP0~@H$HGP0@IELP0fHEHP0HAHP0H$HGP0 H4.HI]H$HD$H1HD$8HD$0E1HD$HD$`A<AHD$XHD$H$HD$HAs<AE1HD$8HD$XHD$0E1HD$HD$HD$`HD$鱿IELP0H|$XH52%HWHH+IM1+ HH$*HD$@1LHH$HF:HH$*Im/ H$H* H$HDŽ$HULbhM%\ID$ H\Ht$(H;5! HOHI H I.IX I M ID$ HL1HHH$w+H$H*? H $H; k!H$HDŽ$HDŽ$HHD$HPH50%HHL*HЅHD$HHD$XE1HD$0HD$E1HD$HD$`HD$ IELP0FH$HGP0H$HGP0H$HGP0`ID$HPH9{HrItHH$HGP0郲HAHP0Hl$PH$HGP0IJHD$`Ld$XE1L|$HH\$L|$pHHD$Ld$XL|$HA=H\$L|$pAH$HD$8E1HD$鏼Ld$XLd$8A=L|$HH\$AL|$pH$MHD$8HD$E1EL|$HL|$pLd$XH\$HDŽ$A=H$A1E1HD$8HD$E1H5*%H=\8%1HH$!H4&H$H*D H4$HDŽ$A :H$HD$HAo1E1HD$8Ht$(HD$XE1HD$0HD$E1HD$HD$`HD$HD$ H$HD$PHD$hH5)%H=x7%1 HH$8!HP%H$H*tH$HDŽ$A;HD$HAHD$(1HD$8HD$XHD$0E1HD$HD$HD$`E1HD$HD$ H$oH $H$A<HD$HAHD$8HD$XHD$0E1HL$(HD$E1HD$HD$`HD$HD$ ιH$HGP0dIEL1E1E1E1P0H $H$A`;HD$HAHD$8HD$XHD$0HL$(HD$HD$HD$`HD$HD$ H$*L&HHgH$HD$HE1HD$8HD$0E1HD$HD$`A<HD$HD$ AHD$XHD$E1飸H5'%H= 5%1 HH$OH"H$H*HDŽ$H$A =HD$HA1E1HD$8HD$XHD$0E1HD$HD$HD$`E1HD$HD$ ߷H$HGP0{IEL1E1E1E1P0H4$H$A;HD$HAHD$8HD$XHD$0Ht$(HD$HD$HD$`HD$HD$ H$;H$HHYH\$H$E1HD$8HD$0E1HD$HD$`A3=HD$ HD$AHD$XHD$E1鸶HX$HH$cH\$A?=AHD$8HD$XE1HD$0HD$E1HD$HD$`HD$HD$ KH$HD$HA<AHD$8HD$XHD$0E1HD$HD$HD$`յH$HGP0Hi#HH$H$A>A1E1HD$8遵H(#HH$PH$A?A1E1HD$8@Ld$XL|$HA=H\$L|$pAH$1E1HD$8HD$E1H"HHH$H$HD$HA:A}HD$(f.HX"HHH$TH$HD$HA:A}HD$(H5"%H=0%1[HI?HImH$HD$HA<A1E1HD$8HD$XE1HD$0HD$E1HD$HD$`HD$HD$ dzI2H4$H$AO:HD$HAs1E1HD$8HD$XHt$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PHD$hHD$@H$HGP0H$HGP0IELP0H$HD$HA<A1HD$8HD$XHD$0E1HD$HD$E1HD$`HD$HD$ nH$HGP0H$HGP0xIFLP0H\$H$AD=AHD$8HD$XHD$0E1HD$HD$HD$`E1HD$HD$ ˱H$1A5?A鰱H$A%?A1HD$8錱H$HGP0H$A7?A1]IEL1E1E1E1P0H$HD$HA<AHD$8HD$XHD$0HD$HD$HD$`HD$HD$ ްHIumI$H5HHFIźLHH$1HD$8AQ?A|Hq!HRH5R0H81HDŽ$H$A>HD$HA1E1HD$8HD$XHD$0E1HD$HD$HD$`E1HD$HD$ ߯H$HD$HA>A1E1HD$8HD$XE1HD$0HD$E1HD$HD$`HD$ sH$HGP0鸮HHI醼LhpMG*I}<*HIp!H$LHHLHH|$AUHuHVHHUHUD$0HR0D$0Ld$XL|$H1H\$0L|$pAr>H\$H$AHD$8HD$E1鄮Ld$XL|$HAp>H\$0L|$pAH\$H$1E1HD$8HD$E18HB@HH$rH >"H=(Dκk'H$H$H$<H5%H=L*%1HI<H)I.8A9AmH$HD$HA:Az1HD$8HD$XHD$0E1HD$(HD$E1HD$HD$`HD$HD$ H$HD$PBH4$H$1HD$HHD$8E1HD$0HD$A9Ht$(HD$`AgHD$HD$ E1H$HD$XHD$HD$PHD$hHD$@|H$HD$HA9Ag1HD$8HD$XHD$0E1HD$(HD$E1HD$HD$`HD$HD$ H$HD$PHD$hHD$@H4$H$A9HD$HAg1HD$8HD$XE1Ht$(HD$0E1HD$HD$HD$`HD$HD$ H$HD$PHD$hHD$@HH4$H$A9HD$HAh1E1HD$8HD$XHt$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PHD$hHD$@駪HB@H0H$}H $H$A9HD$HAh1E1HD$8HD$XHL$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PHD$hHD$@HQ@H8H$jLd$XHD$8E1L|$HH\$A=L|$pH$AHD$E1险Ld$XLd$8A=L|$HH\$AL|$pH$MHD$8HD$E1OHB@H!8H$H|$XLd$XL|$HA=H\$L|$pAH$HD$8E1HD$Ld$XLd$8E1L|$HH\$A=L|$pH$AMHD$8HD$E1饨HQ@H#H$JLd$XL|$HAc>H\$0L|$pAH\$H$1HD$8HD$E1FLd$XLd$8AM>L|$HH\$0AH\$L|$p1MH$E1HD$8HD$E1Ld$XLd$8AK>L|$HH\$0AH\$L|$p1MH$E1HD$8HD$E1鞧Ld$XLd$8A?>L|$HH\$0AH\$L|$p1MH$E1HD$8HD$E1JLd$XL|$HA=H\$L|$pAH$1E1HD$8HD$E1Ld$XL|$HA=H\$L|$pAH$HD$8E1HD$Ld$XL|$H1H\$L|$pA=H$HD$8AHD$E1}Ld$XL|$HA=H\$L|$pAH$1E1HD$8HD$E16Ld$XL|$HA=H\$ L|$pAH\$H$1HD$8HD$E1Ld$XL|$H1H\$L|$pAe=H$HD$8AHD$E1驥Ld$XLd$8A=>L|$HH\$0AH\$L|$p1MH$E1HD$8HD$E1UHB@H:H$ԷH;f! HHHc H$H*= HEHDŽ$HLAHH HAHH$K HA־HIHmٶHEHP0ʶH4$H$A9HD$HAg1HD$8HD$XE1Ht$(HD$0E1HD$HD$HD$`HD$HD$ H$HD$PHD$hHD$@1H鮢HQ@H?@H$gH $H$A9HD$HAh1E1HD$8HD$XHL$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PHD$hHD$@3LnI?H0JcHH4$H$A:HD$HAz1E1HD$8HD$XHt$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$P醢HF0H$HE(H$HE H$HEH$HII~lIIu)M~-H5%HHfH$IMQH$L$H$L$H$MuH5o%HIHH$M~H5%HvHtH$IM~H5a%HQH?H$I)Ld$XL|$HAs=H\$L|$pAH$1HD$8HD$E1Ld$XL|$H1H\$L|$pAq=H$HD$8AHD$E1ϠHZpH?H{?1PHH_Ha!LHHmIk?M6LH|$SI]HHKHIMHD$8IELP0Ht$8ѫLd$XL|$HA=H\$L|$pAH$1E1HD$8HD$E1HB@HsH$ٳLd$XLd$81L|$HH\$0E1H\$L|$pAR>MH$HD$8HD$AE1鎟Ld$XL|$HE1H\$L|$pA=H$HD$8AHD$E1IHQ@Hb>H$SLd$XL|$HA=H\$L|$pAH$E1HD$8HD$E1Ld$XL|$HAz=H\$L|$pAH$HD$8E1HD$鬞Ld$XL|$HAu=H\$L|$pAH$HD$8E1HD$jLd$XLd$8A=L|$HH\$AL|$pH$MHD$8HD$E1 Ld$XLd$8A=L|$HH\$AL|$pH$E1MHD$8HD$E1ӝHLH@HH$Ld$ H$AE<HD$HA1E1HD$8HD$XHD$0E1HD$HD$HD$`E1HD$>H $H$A:HD$HAx1E1HD$8HD$XHL$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PHD$h馜H4$H$A:HD$HAx1E1HD$8HD$XHt$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PLd$XL|$H1H\$0L|$pAa>H\$H$AHD$8HD$E1ΛLd$XLd$8IL|$HH\$A$>L|$pH$AMHD$8HD$E1遛Ld$XL|$HALd$8L|$pHmu HEHP0IMH\$H$A,>A1HD$8HD$E1H$HD$HA:A}HD$(GLd$XL|$HE1Ld$8L|$phH$HGP0Ld$XLd$8E1L|$HH\$A>L|$pH$AMHD$8HD$E1nHxH.H@HH@H$MH4$H$A:HD$HAo1E1HD$8HD$XHt$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PHD$h鰙H$HD$HA;AHD$(HB@HtLH$LMI1H\$H$A,>AHD$8HD$E18HьH$HD$HA =A1E1HD$8HD$XE1HD$0HD$E1HD$HD$`HD$HD$ 鶘H$HD$H1HD$8HD$0E1HD$HD$`A<HD$ AHD$XHD$E1MH$HD$H1HD$8HD$0E1HD$HD$`A<HD$HD$ AHD$XHD$E1ۗH$HD$HA<A1E1HD$8HD$XE1HD$0HD$E1HD$HD$`HD$HD$ fH4$H$A=:HD$HAq1E1HD$8HD$XHt$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PHD$hHD$@ŖHQ@HH$%H$HD$HA<AE1HD$8HD$XHD$0E1HD$HD$E1HD$`HD$HD$ A1HD$8HD$XHD$0E1HD$HD$E1HD$`HD$HD$ 鵔H$HD$HA>A1HD$8HD$XHD$0E1HD$HD$E1HD$`HD$HD$ CH$HD$H1HD$8HD$0E1HD$HD$`A>HD$HD$ AHD$XHD$E1ѓHB@H1H$H|$X4HBHHH$H$H|$H$1E1A'?AwH $H$A:HD$HA}1E1HD$8HD$XHL$(HD$0E1HD$HD$E1HD$`HD$HD$ H$H$HD$HA>AE1HD$8HD$XHD$0E1HD$HD$E1HD$`HD$HD$ ~H4$H$A:HD$HA}1E1HD$8HD$XHt$(HD$0E1HD$HD$E1HD$`HD$HD$ H$H$A>A1E1HD$8ёHB@HtGH$H|$θH$A>A1E1HD$8铑H|$X)IH|$鈸H$HD$HA<AHD$8HD$XHD$0E1HD$HD$HD$`E1HD$HD$ H$HD$HE1HD$8HD$0A<HD$HD$`AHD$HD$ E1HD$XHD$钐H4$H$A{;HD$HA1E1HD$8HD$XHt$(HD$0E1HD$HD$E1HD$`HD$HD$ H$ LTH\$H$A5=AE1HD$8HD$XHD$0E1HD$HD$E1HD$`HD$HD$ 鐏HB@HH$ǖH\$H$E1HD$8HD$0E1HD$HD$`A8=HD$ HD$AHD$XHD$E1 H$H|$H=H$HD$HA;A1HD$(@H4$H$A;HD$HA1HD$8HD$XE1Ht$(HD$0E1HD$HD$HD$`HD$HD$ KHW@HtH$ʉHQ@HtH$HH鰉HIH|$htpHuE鵇QH4$H$A;HD$HA1E1HD$8HD$XHt$(HD$0E1HD$HD$E1HD$`HD$HD$ H$LHQ@HH$C{H $H$1HD$HHD$8E1HD$0HD$A3;HL$(HD$`AHD$HD$ E1H$HD$XHD$鳌HKzH5? %H{ H4$H$AC;HD$HA1E1HD$8HD$XHt$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HB@HH$鎂H $H$AM;HD$HA1E1HD$8HD$XHL$(HD$0E1HD$HD$E1HD$`HD$HD$ H$lH4$H$1HD$HHD$8E1HD$0HD$AO;Ht$(HD$`AHD$HD$ E1H$HD$XHD$H}HQ@HH$ {H4$H$1HD$HHD$8E1HD$0HD$A:Ht$(HD$`A{HD$HD$ E1H$HD$XHD$CHIzHB@HH$߆H$HD$HA;AHD$(1HD$8HD$XHD$0E1HD$HD$HD$`E1HD$HD$ H$麉H:H\A9騼A9靼HB@HtH$tA9逼LtHB@HH$H $H$A.:HD$HAp1E1HD$8HD$XHL$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PHD$hHD$@雈L3>H $H$1HD$HHD$8E1HD$0HD$A;HL$(HD$`AHD$HD$ E1H$HD$XHD$ H $H$A;HD$HA1HD$8HD$XE1HL$(HD$0E1HD$HD$HD$`HD$HD$ H$鈇H4$H$A:HD$HA|1HD$8HD$XE1Ht$(HD$0E1HD$HD$HD$`HD$HD$ H$H$HD$HA:A|HD$(>HB@HH$wH $H$A:HD$HA|1HD$8HD$XE1HL$(HD$0E1HD$HD$HD$`HD$HD$ H$IL&wHW@Ht,H$wH$HD$HA:A|HD$(HwwH4$H$1HD$HHD$8E1HD$0HD$A:Ht$(HD$`A|HD$HD$ E1H$HD$XHD$wH $H$A:HD$HA{1HD$8HD$XE1HL$(HD$0E1HD$HD$HD$`HD$HD$ H$HPHE!H H5 L|$HLd$XL|$pH81"kHUHfnH$LpH5U$LHE/9LIgH$A ?A1E1HD$8AH$A ?A1E1HD$8HB@Ht5H$H|$H$A>A1E1HD$8܃H|$rH$HD$HA;A}HD$(H $H$A:HD$HA|1E1HD$8HD$XHL$(HD$0E1HD$HD$E1HD$`HD$HD$ H$$H$HD$HA;A}HD$(H4$HD$H1HD$8HD$0E1HD$HD$`A:Ht$(HD$A}HD$ H$E1HD$XHD$醂H $H$A;HD$HA}1HD$8HD$XE1HL$(HD$0E1HD$HD$HD$`HD$HD$ H$H4$H$A;HD$HA}1HD$8HD$XE1Ht$(HD$0E1HD$HD$HD$`HD$HD$ H$送H $H$1HD$HHD$8E1HD$0HD$A;HL$(HD$`A}HD$HD$ E1H$HD$XHD$HB@HH$;uH4$H$1HD$HHD$8E1HD$0HD$A;Ht$(HD$`AHD$HD$ E1H$HD$XHD$dH4$H$A:HD$HAz1HD$8HD$XE1Ht$(HD$0E1HD$HD$HD$`HD$HD$ H$HD$PH $H$1HD$HHD$8E1HD$0HD$A:HL$(HD$`AzHD$HD$ E1H$HD$XHD$HD$PLH $H$1HD$HHD$8E1HD$0HD$A:HL$(HD$`AzHD$HD$ E1H$HD$XHD$HD$P~H $H$1HD$HHD$8E1HD$0HD$A:HL$(HD$`AzHD$HD$ E1H$HD$XHD$HD$P4~HB@HH$hmH4$H$A:HD$HAz1HD$8HD$XE1Ht$(HD$0E1HD$HD$HD$`HD$HD$ H$HD$P}L*lHQ@HH$kH $H$AM:HD$HAs1E1HD$8HD$XHL$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PHD$hHD$@|HfIMkHzH=!H8HyH$HD$HE1HD$8HD$0E1HD$HD$`A6<HD$HD$ AHD$XHD$E1|H$HD$HA8<AE1HD$8HD$XHD$0E1HD$HD$E1HD$`HD$HD$ {Ld$ HDŽ$AE<H$HD$HA1E1HD$8HD$XHD$0E1HD$HD$E1HD$`HD$-{HQ@HH$8sH4$H$1HD$HHD$8E1HD$0HD$A;Ht$(HD$`AHD$HD$ E1H$HD$XHD$zH,IrH$HD$HA;AHD$(bInHD$HA<AɯHB@HtgH$H|$XLd$XL|$HAf>H\$0L|$pAH\$H$1E1HD$8HD$E1yLocH|$X`魙H4$H$1HD$HHD$8E1HD$0HD$A\;Ht$(HD$`AHD$HD$ E1H$HD$XHD$8yHB@HH$~H$HD$HE1HD$8HD$0A<HD$HD$`AHD$HD$ E1HD$XHD$xH$HD$HE1HD$8HD$0A<HD$HD$`AHD$HD$ E1HD$XHD$BxHI}H $H$1HD$HHD$8E1HD$0HD$A;HL$(HD$`AHD$HD$ E1H$HD$XHD$wHB@HH$pHB@HH$hH4$H$A:HD$HAn1E1HD$8HD$XHt$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PHD$hvLNoHvIgHB@HH$nH $H$1HD$HHD$8E1HD$0HD$Al;HL$(HD$`AHD$HD$ E1H$HD$XHD$5vHIvmH4$H$An;HD$HA1HD$8HD$XE1Ht$(HD$0E1HD$HD$HD$`HD$HD$ H$uH $H$1HD$HHD$8E1HD$0HD$A<HD$`HD$AHD$ HL$(E1HD$XHD$'uIFLP0A9AmH5!HT$8H8HHE HT$8cH\$H$AA=AHD$8HD$XHD$0E1HD$HD$HD$`E1HD$HD$ qtHW@HH$B|H$HD$H1HD$8HD$0E1HD$HD$`A<HD$HD$ AHD$XHD$E1sH $H$1HD$HHD$8E1HD$0HD$A;HL$(HD$`AHD$HD$ E1H$HD$XHD$fsH$1E1HD$8HD$0E1HD$HD$`A%=HD$ AHD$XHD$E1sH\$H$AF=AHD$8HD$XHD$0E1HD$HD$HD$`E1HD$HD$ rH/vzH$A#=A1E1HD$8HD$XHD$0E1HD$HD$E1HD$`HD$HD$ rHB@HtBH$xH$A?A1E1HD$8qHQ@HtH$ךHkzxH^šH$A>A1E1HD$8qH$A!?A1E1HD$8kqH$1HD$8A#?AGqHB@HpH$隚HIYH|$XIzH;~A9AmA9AlH|$8zH $H$A:HD$HAn1E1HD$8HD$XHL$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PHD$h=pH $H$A2:HD$HAp1E1HD$8HD$XHL$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PHD$hHD$@oHQ@HtH$ZA9բHZH4$H$E1HD$HHD$8A;HD$0HD$AHt$(HD$`E1HD$HD$ HD$XHD$nH $H$1HD$HHD$8E1HD$0HD$A;HL$(HD$`AHD$HD$ E1HD$XHD$~nA9锡H4$H$A0:HD$HAp1E1HD$8HD$XHt$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PHD$hHD$@mH4$H$A:HD$HAz1HD$8HD$XE1Ht$(HD$0E1HD$HD$HD$`HD$HD$ H$HD$PFmHB@HH$kH4$H$A;HD$HA1E1HD$8HD$XHt$(HD$0E1HD$HD$E1HD$`HD$HD$ H$lH-!H8t%肼HyHB@HtZH$H|$`YuLd$XL|$HA=H\$ L|$pAH\$H$1E1HD$8HD$E1&lH|$`輿uHB@HtH$Z_HB@HtH$^莿@_脿@^H¹HHߊH,!HT$0H8M虻HT$0頀HLd$XL|$HLd$8L|$pHyHMH\$H$A>A1E1HD$8HD$E1AkLپ}@苿MH\$H$A>A1E1HD$8HD$E1jHy+!H5"M1E1E1H81(H\$H$A>AHD$8HD$jH$A>A1E1HD$8qjH$A>A1E1HD$8JjH$A>A1E1HD$8#jH+!H8dwHB@Ht0H$LH$A ?A1E1HD$8iLaH $H$Af:HD$HAu1E1HD$8HD$XHL$(HD$0E1HD$HD$E1HD$`HD$HD$ H$HD$PHD$h$iL輼gH\$H$E1HD$8HD$0A==HD$HD$`AHD$ HD$E1HD$XHD$hH$HD$HE1HD$8HD$0A<HD$HD$`AHD$XHD$PhH$HD$HA<AHD$8HD$XHD$0E1HD$HD$HD$`gH$HD$HE1HD$8HD$0A<HD$HD$`AHD$XHD$gHD$HA\<AHD$8HD$XE1HD$0HD$E1HD$HD$`HD$cgH$HD$HAW<AE1HD$8HD$XHD$0E1HD$HD$E1HD$`HD$fH$HD$HE1HD$8HD$0E1HD$HD$`AU<HD$AHD$XHD$E1gfHB@HH$H|$X{HB@HH$6H$HD$HE1HD$8HD$0E1HD$HD$`At<AHD$XHD$eH$HD$HAl<A1E1HD$8HD$XE1HD$0HD$HD$HD$`HD$oeHD$HAj<A1HD$8HD$XHD$0E1HD$HD$HD$`HD$9eH$HD$HE1HD$8HD$0E1HD$HD$`Aq<HD$AHD$XHD$dH|$X@H̟Ld$ H$1HD$HE1HD$8HD$0E1HD$HD$`HD$AG<AHD$XHD$E1'dHB@HtqH$H|$ HD$HA<AHD$8HD$XE1HD$0HD$HD$HD$`cLUH鵠H|$ C鄟H6)QM fH\$H$1E1HD$8HD$0E1HD$ HD$AW=AHD$XHD$E1*cH\$H$AU=A1E1HD$8HD$XE1HD$0HD$E1HD$HD$HD$ bLjpM9I}.H#!Ht$(HHI*HHAUI $HHQHI$ID$LP0ߣLd$XL|$HA=H\$L|$pAH$1E1HD$8HD$E1bH4$H$1HD$HHD$8E1HD$0HD$A ;Ht$(HD$`A}HD$HD$ E1H$HD$XHD$aH$HD$HA<A1E1HD$8HD$XE1HD$0HD$E1HD$HD$`HD$HD$ aHEHP0H!!HRH5L|$HLd$XL|$pH81?餤HrHlLbI+f.AWHAAVAUATUSHhH\$pH$HD$XdH%(H$X1HHD$`HD$hկ`H/H1H路$8D$p=L$LH_IHپL1z1L+L-AA11HD$XHtHHQHHHD$`HtHHQHHHD$hHtHHQHHHt HmHt H+H=6$H=LDDCH=$H/uHGP0H$H$XdH3%(f_Hh[]A\A]A^A_$8D$rfD1YHH$^H=>1HHi$^H5$H=c11AHH\$HHұHH8$^HH=0"HH$^H=$H5H<H-]$HHM$)H{诬HEHH8t:H(H+Hz{ ul{"uHCH{HpHEHH8uH[$M$HH>$AL-A11f.HCH{1Hp}HEY@AL-FA11X@H|$XHGP0`H|$`HGP0eH|$hHGP0jHEHP0dfHCHP0H=$`裭HH!H5]H8ĪkAL-~A11@1HH$wHH$vͮHHk$]賮HHI$]虮HH'$@]H}HH$\ cHH$\IHH${\/HH$:\pHH{$[HǀHHW$ ߭HH5$ʉŭHH$H詭HH$FH=11 HH$H=11HH$H=11ɧHHw$qH-HHS$.H=11荧HH+$H=i11mHH$H=^11MHH$=H!t"H$H=$H58#H=G$rHH8$TH=$VHH$FTH=$:HH$TH=$HH$SH=$HH$=TH=$HH$HTH=$HH`$STH= $HH<$VH!1HHHHH$v1HHH޿转HH$*1HHH޿蘬HH$ރH5$1uHH$1HHH޿PHHV$1HHH޿+HH)$ˆ1HHH޿HH$vH5$1HH$,1HHH޿辫HH$1HHH޿虫HHg$1HHH޿tHH:$HH5$1QHH$1HHH޿,HH$Z1HHH޿HH$1HHH޿HH$‰1HHH޿轪HH[$vH5$1蚪HH0$,1HHH޿uHH$1HHH޿PHH$1HHH޿+HH$HH5D$1HH~$1HHH޿HHQ$1HHH޿辩HH$$f1HHH޿虩HH$H5$1vHH$І1HHH޿QHH$1HHH޿,HHr$81HHH޿HHE$H5 $1HH$ґ1HHH޿迨HH$1HHH޿蚨HH$:H5s$1wHH$H50$1THHj$1HHH޿/HH=$ZH5$1 HH$H5$1HH$ƏH5:$1ƧHH$|1HHH޿衧HH$01HHH޿|HHb$H5$1YHH7$H5$H=$H蓦HH$MH5d$H=e$HmHH$H5$1HH$1HHH޿ŦHH$jH5$1袦HHh$ 1HHH޿}HH;$Ԍ1HHH޿XHH$H59$15HH$>H5$1HH$H5$1HH$H5$1̥HHb$`H5$1詥HH7$H5b$1膥HH $̊H57$1cHH$H5 $1@HH$8H5$1HH$H5$1HH`$H5K$1פHH5$ZH5$1贤HH $H5E$1葤HH߾$ƈH5 $1nHH$|H5$1KHH$2H5$1(HH^$H5$1HH3$H5v$1HH$TH5K$1迣HHݽ$ H58$1蜣HH$H5$1yHH$vH5$1VHHL$,H5o$13HH!$H5L$1HH$H5$1HH˼$NH5$1ʢHH$H5C$1觢HHu$H5$1脢HHJ$pH5$1aHH$&H5r$1>HH$܃H5G$1HHɻ$H5$1HH$HH5$1աHHs$H5$1財HHH$H5$1菡HH$jH5$1lHH$ H5e$1IHHǺ$ցH5:$1&HH$H5$1HHq$BH5$1HHF$H5$1轠HH$H5$1蚠HH$dH5$1wHHŹ$H5h$1THH$H5E$11HHo$H5$1HHD$wH5$1HH$vH5\$1ȟHH$rvH59$1襟HHø$(vH5$1肟HH$uH5$1_HHm$uH5$1$qH5$1uHH$qH56$1RHH$>qH5#$1/HH$pH5$1 HH$pH5$1HHg$`pH5$1ƜHH<$pH5$1補HH$oH5T$1耜HH$oH5)$1]HH$8oH5$1:HH$nH5$1HHe$nH5$1HH:$ZnH5}$1ћHH$nH5r$1讛HH$mH5G$1苛HH$|mH5$1hHH$2mH5$1EHHc$lH5$1"HH8$lH5$1HH $TlH5$1ܚHH$ lH5u$1蹚HH$kH5J$1薚HH$vkH5$1sHHa$,kH5$1PHH6$jH5$1-HH $jH5$1 HH$NjH5k$1HH$jH5H$1ęHH$iH5$1衙HH_$piH5$1~HH4$&iH5$1[HH $hH5$18HHް$hH5$1HH$HhH5V$1HH$gH5+$1ϘHH]$gH5$1謘HH2$jgH5$1艘HH$ gH5ʿ$1fHHܯ$fH5$1CHH$fH5t$1 HH$BfH5I$1HH[$eH5$1ڗHH0$eH5$1跗HH$deH5ؾ$1蔗HHڮ$eH5$1qHH$dH5z$1NHH$]H5O$1+HHY$Z]HHHmHH$]HHHOHH$\HH1͖HH$\H5$1誖HHȭ$=\1HHH޿腖HH$[1HHH޿`HHn$[1HHH޿;HHA$Y[1HHH޿HH$ [1HHH޿HH$ZH R$H$1H5"$HH$iZHI$L r$1La$D$0@HD$HT$(Hv$LD$8LL$LL$L $HT$ 17HH$YH`$L $ Lճ$H ~$H'$H5@$HD$(H $HD$ Hp$HD$Hd$HD$H$HD$H$H$1єHH$,YH$L $1Lr$D$0V HD$HT$(H$LD$8LL$LL$L $HT$ 1HHH$XHq$L ʳ$ L$H $H8$H5Q$HD$(H$HD$ H$HD$Hu$HD$H!$HD$H$H$1HH$WH$L $1L$D$0q HD$HT$(H$LD$8LL$LL$L $HT$ 1YHH$_WH$L ۲$ L$H $HI$H5b$HD$(H.$HD$ H$HD$H$HD$H2$HD$H$H$1HHɩ$VHĵ$L $1L$D$0 HD$HT$(H$LD$8LL$LL$L $HT$ 1jHH$"VH$L $ L$H $HZ$H5s$HD$(H?$HD$ H$HD$H$HD$HC$HD$H$H$1HHҨ$uUHʹ$L $1L$D$0 HD$HT$(H$LD$8LL$LL$L $HT$ 1{HH $TH$L $ L$H ¯$Hk$H5$HD$(HP$HD$ H$HD$H$HD$HT$HD$HЯ$H$1HHۧ$8THֳ$L ǹ$1L$D$0 HD$HT$(H˷$LD$8LL$LL$L $HT$ 1茊HH$SH$L $ L*$H Ӯ$H|$H5$HD$(Ha$HD$ HŲ$HD$H$HD$He$HD$H$H$1&HH$RH$L ظ$1LǸ$D$0 HD$HT$(Hܶ$LD$8LL$LL$L $HT$ 1蝉HH$kRHƭ$L $ L;$H $H$H5$HD$(Hr$HD$ Hֱ$HD$Hʮ$HD$Hv$HD$H$H$17HH$QH$L $1Lط$D$0 HD$HT$(H$LD$8LL$LL$L $HT$ 1讈HH$$.QH׬$L 0$ LL$H $H$H5$HD$(H$HD$ H$HD$Hۭ$HD$H$HD$H$H$1HHH$PH$L $1L$D$0* HD$HT$(H$LD$8LL$LL$L $HT$ 1过HH-$OH$L A$ L]$H $H$H5ȭ$HD$(H$HD$ H$HD$H$HD$H$HD$H$H$1YHH$GOH$L $1L$D$0E HD$HT$(H$LD$8LL$LL$L $HT$ 1ІHH6$NH5a$1͌HHk$YNH5>$1誌HH@$TH=p`HH$rTH=ȳؠHH$5TH=H 跠HHŴ$SH=0薠HH$meHW$H=x$H$HJH;$螇H=$11H-L$HH%H5-HHHHHH2H=w$H$H5k蜄Hݕ$H=H׳$zHH,7H5ݲHH+HHfH* H9E H H5ҺH8BHmHP H;H5AL-֪Ap11HHHu HCHP0AL-Ap11fAÁL-Ap11HCHP0DAL-VA11hHCHP0HEHP0Ht HgHMH$HQHHU4Ha= HIJ$4 H$C6HP6H="$1HHHD$XPH5$H=T$O//HT$XH*$H讁HHD$h.H+u HCHP0HT$hH5 $H=}$xo6.HT$hH*u H|$hHGP0H=j$HD$h脊HHD$h-H5G$H'HH-HT$hH*u H|$hHGP0H5$H=$HHD$hnS-H+u HCHP0H=ޒ$HH-H5$H覀HHD$h,H+u HCHP0HT$hH5ԍ$H=u$pn,HT$hH*u H|$hHGP0H=b$HD$h|HHD$h8,H5$HHH,HT$hH*u H|$hHGP0H5M$H=$HHD$hm+H+u HCHP0H=֑$HHi+H5&$HHHD$h2+H+u HCHP0HT$hH5$H=m$hm*HT$hH*u H|$hHGP0H=Z$HD$htHHD$h*H5$HHH[*HT$hH*u H|$hHGP0H5$H=$HHD$hl*H+u HCHP0H=ΐ$HH)H5Ώ$H~HHD$h)H+u HCHP0HT$hH5$H=e$`l>)HT$hH*u H|$hHGP0H=R$HD$hlHHD$h(H5'$H~HH(HT$hH*u H|$hHGP0H5$H=ޕ$HHD$hk[(H+u HCHP0H=Ə$HH(H5$H}HHD$h'H+u HCHP0HT$hH5Ċ$H=]$Xk'HT$hH*u H|$hHGP0H=J$HD$hdHHD$h@'H5W$H}HH 'HT$hH*u H|$hHGP0H5%$H=֔$HHD$hj&H+u HCHP0H=$HHq&H5N$H|HHD$h:&H+u HCHP0HT$hH5$H=U$Pj%HT$hH*u H|$hHGP0H=B$HD$h\HHD$h%H5$H{HHc%HT$hH*u H|$hHGP0H5U$H=Γ$HHD$hi %H+u HCHP0H=$لHH$H5&$H~{HHD$h$H+u HCHP0HT$hH5$H=M$HiF$HT$hH*u H|$hHGP0H=:$HD$hTHHD$h#H5_$HzHH#HT$hH*u H|$hHGP0H5-$H=ƒ$HHD$hhc#H+u HCHP0H=$уHH!#H5$HvzHHD$h"H+u HCHP0HT$hH5$H=E$@h"HT$hH*u H|$hHGP0H=2$HD$hLHHD$hH"H5'$HyHH"HT$hH*u H|$hHGP0H5$H=$HHD$hg!H+u HCHP0H=$ɂHHy!H5ވ$HnyHHD$hB!H+u HCHP0HT$hH5$H==$8g HT$hH*u H|$hHGP0H=*$HD$hDHHD$h H5O$HxHHk HT$hH*u H|$hHGP0H5$H=$HHD$hf H+u HCHP0H=$HHH5v$HfxHHD$hH+u HCHP0HT$hH5D$H=5$0fNHT$hH*u H|$hHGP0H="$HD$h$HVvHHD$hJ&H+u HCHP0HT$hH5 $H=%$ d%HT$hH*u H|$hHGP0H=$HD$h,HHD$h%H5G$HuHHs%HT$hH*u H|$hHGP0H5$H=$HHD$hc%H+u HCHP0H=$~HH$H5$HNuHHD$h$H+u HCHP0HT$hH5$H=$cV$HT$hH*u H|$hHGP0H= $HD$h$~HHD$h$H5$HtHH#HT$hH*u H|$hHGP0H5$H=$HHD$hbs#H+u HCHP0H=~$}HH1#H5n$HFtHHD$h"H+u HCHP0HT$hH5<$H=$b"HT$hH*u H|$hHGP0H=$HD$h}HHD$hX"H5$HsHH#"HT$hH*u H|$hHGP0H5$H=$HHD$h}a!H+u HCHP0H=v$|HH!H5V$H>sHHD$hR!H+u HCHP0HT$hH5$$H= $a!HT$hH*u H|$hHGP0H=$HD$h|HHD$h H5ׁ$HrHH{ HT$hH*u H|$hHGP0H5$H=$HHD$hu`# H+u HCHP0H=n${HHH5$H6rHHD$hH+u HCHP0HT$hH5Ԃ$H=$`^HT$hH*u H|$hHGP0H=$HD$h {HHD$h$(H5G$HqHH'HT$hH*u H|$hHGP0H5$H=~$HHD$hm_'H+u HCHP0H=f$zHHU'H5$H.qHHD$h(H+u HCHP0HT$hH5|$H=$^y(HT$hH*u H|$hHGP0HD$h]HHD$h0-H$H5$H^,H$H5$H|$h^,HX$H5Y$H|$hg^~,H($H5)$H|$hG^B,H$H5$H|$h'^,Hȅ$H5Ʌ$H|$h^+H$H5$H|$h]+Hh$H5i$H|$h]R+H8$H59$H|$h]+H$H5 $H|$h]*H؄$H5ل$H|$hg]*H$H5$H|$hG]b*Hx$H5y$H|$h']&*HH$H5I$H|$h])H$H5$H|$h\)H$H5$H|$h\r)H$H5$H|$h\6)H$H5$H|$h\(HX$H5Y$H|$hg\(H($H5)$H|$hG\(H$H5$H|$h'\F(HȂ$H5ɂ$H|$h\ (H$H5$H|$h['Hh$H5i$H|$h['H8$H59$H|$h[V'H$H5 $H|$h['H؁$H5ف$H|$hg[&H$H5$H|$hG[&Hx$H5y$H|$h'[f&HH$H5I$H|$h[*&H$H5$H|$hZ%H$H5$H|$hZ%H$H5$H|$hZv%H$H5$H|$hZ:%HP$H5Q$H|$hgZ$H $H5!$H|$hGZ$H$H5$H|$h'Z$H$H5$H|$hZJ$H$H5$H|$hY$H`$H5a$H|$hY#H0$H51$H|$hY#H$H5$H|$hYZ#H~$H5~$H|$hgY#HT$hH5s}$H=L$GY"HT$hH*zA11PH|$XHGP0AL-zA11#H|$XHGP0AL-yA11Hyu$g$yHX$AL-yA11鸥Hy7$F)$yH$Hly$$yH$HHy$$yHҁ$uH!yȁ$$yH$NHx$q$yH$'Hxz$l$yH]$H|$XHGP0H|$XHGP0nAHD$XHtHHQHHHD$`HD$XHtHHQHHpH=$HD$``S!H $xH=D iHT$hHt$XH|$` eOHH0H${$HHHEH{$HH=8y$tHH0Hmu HEHP0H5z$HpHHf0H5z$H==$H5VHmu HEHP0H+u HCHP0HT$`H*u H|$`HGP0HT$XHD$`H*u H|$XHGP0HT$hHD$XH*u H|$hHGP0H|$@LLHD$hLa^H|$`HGP0H|$XHGP0AaL-vA@11魢IGLP0IELP0HAHP0H|$hHGP01AmL-@vAV11RH*v~$~$yH~$WA11H|$@LLL-uf`HEHP0H~$H|$hHGP0AyL-uAq11¡H|$hHGP0&AL-uA11镡H|$hHGP0RAL-VuA11hAL-:uA11LH5 Hq H81GTHP A,A11H|$XHGP0jH|$`HGP0A͂L-tAE11̠AL-tA*11鰠AL-tA11锠AL-ftA11xH|$hHGP0ABA5H{ H5H;LNAL- tA11AL-sA11AL-sA11AL-sA11˟MH5KH H81R~ H5THhs|$|$RH{$鯡HAs{${$QH{$鈡Hs{${$PH{$aHr{${$OH}{$:Hrs{$e{$NHV{$HrL{$>{$MH/{$H~r%{${$LH{$ŠHWrz$z$KHz$鞠AăL-*rA1>AƒL-rA11"AAL-q11#AL-qA11AL-qA11ΝAL-qA11鲝AL-qA11閝AL-hqA11zAL-LqAc11^AL-0qAc11BAL-qAl11&AAlAL-pAc11AAkAL-pAk11̜AxL-pAc11鰜AmL-pAj11锜AkAjmAaL-UpAc11gAVAi@ATL-(pAi11:AJL- pAc11A?L-oAh11A=AhA3L-oAc11՛A(AgA&L-oAg11騛AL-zoAc11錛AL-^oAf11pAAfIH|$XHGP0AL- oAc112AeL-oA11AbL-nA1A`L-nA1AVL-nA1ȚASL-nA1鮚AQL-nA11钚AGL-dnA11vADL-HnA1\ABL-.nA1BA8L-nA1(A5L-mA1A3L-mA11A)L-mA11֙A&L-mA1鼙A$L-mA1颙AL-tmA1鈙AL-ZmA1nAL-@mA11RA L-$mA116AL-mA1AL-lA1AL-lA1AL-lA1ΘAL-lA11鲘AL-lA11閘AL-hlA1|AL-NlA1bAބL-4lA1HAۄL-lA1.AلL-lA11AτL-kA11ĀL-kA1ܗAʄL-kA1—AL-kA1騗AL-zkA1鎗AL-`kA11rAL-DkA11VAL-(kA1jAr 1RAUL-$jA 18APL- jA 1ACL-iA 1A>L-iA 1A1L-iAh1ЕA$L-iA1鶕AL-iA1霕AL-niA1邕AL-TiA1hAL-:iA1NA݃L- iA14AكL-iA1AփL-hA1A҃L-hA1A̓L-hA1̔AʃL-hA1鲔AǃL-hA1阔AL-jhA1~AL-PhA1dAL-6hA1JAL-hA11.AL-hA11AL-gA1AL-gA1ޓAL-gA1ēAL-gA1骓A}L-|gA11鎓AsL-`gA11rApL-DgA1XAnL-*gA1>AdL-gA1$AaL-fA1 A_L-fA11AUL-fA11ҒARL-fA1鸒APL-fA1鞒AFL-pfA1鄒ACL-VfA1jAAL-AtL-bA1$AqL-aA1 AoL-aA11AEL-aA 11ҍABL-aA 1鸍A@L-aA 1鞍A6L-paA 1鄍A3L-VaA 1jA1L-HQ3Z$z%Z$HZ$HeQ Z$xY$HY$H>QY$mY$HY$HQY$kY$HY$HPY$iY$HzY${HPpY$gbY$HSY$THPIY$;Y$H,Y$-H{P"Y$Y$HY$HTPX$X$HX$H-PX$X$HX$HPX$X$HX$HOX$xX$HiX$jHO_X$hQX$uHBX$CHO8X$f*X$jHX$HjOX$^X$_HW$HCOW$\W$THW$HOW$W$IHW$HNW$W$>HW$HNuW$gW$3HXW$YHNNW$@W$(H1W$2HN'W$W$H W$ HYNW$V$HV$H2NV$V$HV$H NV$V$~HV$HMV$}V$~HnV$oHMdV$VV$~HGV$HHM=V$/V$~H V$!HoMV$V$~HU$HHMU$U$~HU$H!MU$@U$~HU$HLU$>U$~HU$HLzU$<lU$~H]U$^HLSU$2EU$~H6U$7HL,U$0U$~HU$H^LU$.T$~HT$H7LT$T$x~HT$HLT$T$m~HT$HKT$T$b~HsT$tHKiT$[T$W~HLT$MHKBT$4T$L~H%T$&HtKT$ T$A~HS$HMKS$FS$6~HS$H&KS$=S$+~HS$HJS$ S$ ~HS$HJS$ qS$~HbS$cHJXS$n JS$ ~H;S$HC3L$%L$yHL$HeC L$K$yHK$H>CK$K$yHK$HCK$#K$zHK$HBK$K$zHzK${HBpK$bK$zHSK$THBIK$;K$zH,K$-H{B"K$K$zHK$HTBJ$J$zHJ$H-BJ$J$zHJ$HBJ$J$yzHJ$HAJ$xJ$nzHiJ$jHA_J$QJ$czHBJ$CHA8J$*J$XzHJ$HjAJ$J$MzHI$HCAI$~I$BzHI$HAI$lI$7zHI$H@I$ZI$,zHI$H@uI$EgI$!zHXI$YH@NI$ @I$}H1I$2H@'I$ I${}H I$ HY@I$> H$p}HH$H2@H$< H$e}HH$H @H$3 H$Z}HH$H?H$1 }H$O}HnH$oH?dH$ VH$D}HGH$HH?=H$ /H$9}H H$!Ho?H$ H$.}HG$HH?G$ G$#}HG$H!?G$ G$}HG$H>G$} G$ }HG$H>zG${ lG$}H]G$^H>SG$y EG$|H6G$7H>,G$, G$|HG$H^>G$* F$|HF$H7>F$! F$|HF$H>F$ F$|HF$H=F$F$|HsF$tH=iF$[F$|HLF$MH=BF$4F$|H%F$&Ht=F$ F$|HE$HM=E$dE$|HE$H&=E$\E$|HE$H<E$E$~|HE$H<E$qE$s|HbE$cH<XE$JE$h|H;E$B${H/B$0H~9%B$B${HB$ HW9A$<A$v{HA$H09A$;A$k{HA$H 9A$4A$`{HA$H8A$/{A$U{HlA$mH8bA$.TA$J{HEA$FH8;A$-A$?{HA$Hm8A$A$4{H@$HF8@$@$){H@$H8@$@${H@$H7@$@${H@$H7x@$j@${H[@$\H7Q@$C@$zH4@$5H7*@$@$zH @$H\7@$?$zH?$H57?$-?$zH?$H7?$,?$zH?$f.fHHTHH1HielHH=qudžpHdždždžÐHc=ptBPHljHH H1HH%V,H1HHH1HHH1ÐHH HBH%H HHH3` H%߰H1HJuHH HBH%H HHH3H%߰H1HJouHHxHH HʃHH3` Hف߰H1Hx@AWAVAUATUHSH(HH|$HT$HL$u;HHHH~@HL$HHHH9uH([]A\A]A^A_HHH HHH HHH HHH HHH II I H|$~E1LH9EwJf.HXIM!L9rLd$HD$N$IL;|$ZH9vfDHIHI I M!L9rff.AWIAVAAUATUSHHT$u,HI~E7II9uH[]A\A]A^A_É  Љ ‰ Љ H|$~ME1DLh!9rDCIL;l$uH[]A\A]A^A_ÐAWAVAUATAUSHfu&HHQ~fD!HH9uH[]A\A]A^A_LAAfAA DfD ljf AfAA H~HL,Q1D1!fA9suH D!fA9rDHfSL9uH[]A\A]A^A_DAWAVAUATAUSH@u%HH~ D!HH9uH[]A\A]A^A_ÉA@ @ AAA H~L,LH1fDD!A8suH D!A8rDHKL9uH[]A\A]A^A_@t[ATMUH,SH1H,fD[]A\f.HSH9tڅuL^ H~AHDD HH9ufUSHH" HH HHH H [H]HgHHDf.AT1HUSHtbHIHH HHH HHH HHH HHH HH H ŸH9w"f.Lx H!H9r[]A\L H!H9sL H!H9rfDUSHH2 HHH# H*HH*Y=JXY9JH[]f.AVHIAUIATIUSvNHHLHH HˆCHSHHHSHCwIFAHMdMu[]A\A]A^DLMm DIAD$HM9u[]A\A]A^f.ATIH5"IUHStCH=I HHtCHLHdHH[1]A\DH=Hl HHu[]A\f.AV1ҾAUATUSHH dH%(HD$1U u{H ǃpHǃǃ1ǃH fHHHH=u1HL$dH3 %(-H []A\A]A^1H'HHPHH)HH H1HHHH1HHH H)H$HPHH)HH H1HIII1LIH I)HD$MIHPHH)HH H1HIII1LIH I)HPHLHH)L1IHL1H L1H1H1HHHHH1HHH H)H1HH1H1P ATIUHSHH tLHH߉D$ TD$ H[]A\SHH f.H f(HX\EFT$ f(T$Xf(Y\ Ff(YXf.FsfWf.ztf(L$T$\$ YGF\$T$L$^Qf.zJYf(ǃYH [LJHLJH [L$T$ L$f(T$f.H+@LDLHL1HielH HH=puUHpHME11SLJpL LJHIN\NMIM1Mi fM3 AHoM ~HxHL9LNHuo@HLDH4LHL1HieX]H3)HHoH~HxHHuHLJLJLJH[]fUf(SH(-Df.f(\H,H*Xf(CHDH$L$f(d$^\EDYl$X +Df.X\rAH~<\CL$Hf($L$H9$\~H(f([]Ðf(1DHL$$\L$YX$HHG B\f(R DHfWÐHD$AYD$HfDHL$$L$YX$HSHH0f.BD$%Af.d$|$B\=BY|$(Qf.=A^|$HHL$fWYX Af.sf(HD$YYL$T$ vBf(5RAYYY\f.wZT$ D$D$T$  BYYA\T$XYT$(Xf.L$2D$(YD$H0[D$@L$ \^D$}t$L$ f(D$Yf( @^\T$\$\f.sH HD$. F@T$\L$f.` (@D$f(^L$#\$f.rH0[H0[<f(!HL$1L$HYfSHH0-?D$f.L$s>D$HD$HD$L$H0[X^f(@f.rf.HHD$  2?D$(^L$D$ + ?D$^L$D$( XD$%>f.rf.?v|$H0[^f(f.D$ f(D$(^T$T$jf(T$^L$f(_L$ \\$f(T$3\$L$ D$\f(XD$ T$H0[\f(HY ?oHXfDSf(HH\$ $ $Hf(f(Y$\$$HY[^f(fAWAVAUATIUHSH$t H9' a=L$Džf(\f.t$Hf(\|$@I*L$H|$@T$YX\$f(L$\$H,Y\$@L$T$Qf.\$pHG \$@Y=T$Y=L$\f(H*=Q=L$X|$(T$fD(t$f(XXP=f($f(57=AXA\^f(l$x\d$0X5=fD(t$Pt$HYf(f(\^f(Yf(f(fA(XYf(D$`f(^\L$@Yf(fA(^YXYfA(AX^D$hXAYXT$Xl$8t$ MI)IEH$HxL$ HYL$`L$f(f.L$2f.L$nt$f(|$P\H*f(i:Y^XXD$0\X\$(fTU:^\f.cT$Xd$RL,d$T$XMI)LH?HL1H)H~D$pH*YD$(\f.\$HID$I9^\$@H*Y[HSI9|,f.H*f(HI9^f(\Y}f.@$M)f.|$(MGH[]A\A]A^LA_|$H|$@$|$H Df.L$8L$XT$wT^D$`XD$0&L,ML$XT$f(\d$Y8Yd$`^D$h|$x\f(L,T$L$XM9f(\d$8Yj8Yd$hNIGH9DH*f(HH9^f(\^~f.Df(H|$pH^89f(H^X,9YH*X#9^XD$(Yf(XT$X^f(L$lL$f(T$Xf(\f. Xf.IGH*$fD($H*HCDY|$XL*ID$L)D$H*$fA(D$fE(^f(f(l$YEYY$D$$|$X$f($^D$l|$$Y|$HD$Xd$@Yf(^:I*$D-7Y$D$D%7D7fA(D7D$D$$Xd$($$$Yd$XXI*YfA(A^XfA(\f(fA(A^\f(A^\7f(A^\f(fA(A^D 6A^XfA(A^\f(fA(A^\f(fA(A^\f(f(A^\f(fA(^A^XfA(^\f(fA(^\f(fA(^\f(f(^\f($^D^A^E\D^XE\D^E\D^A\^D$A^Xf.@YT$$\X|L,f.Hf(|$Hf(|$@|$$|$0|$x|$P|$`|$h|$|$8|$ H*YY|$p=Q4|$(kf(T$T$L$@f.ATf(IUSHH@t H9I*L%2ǃ\\$ f(d$L$zL$Yk\$ L$f(d$D$Yf(YX5[2Qf.f(e=3YXf.H,HH\$d$t$1d$f.f(\$w6kL\H)HH*YYH*Y^f.v7HPH9}H\$d$NL$1d$f.\$wH@[]A\f.f(zLYXH,2@f.SMH|$f(|$(\$ T$d$Q|$(\$ T$d$vf(l$8\$0T$(L$ t$ l$8\$0T$(L$ d$t$K@ 1SHf.rH* 2Yf.r?[ X01\H*Yf.f(sDH)H[@[2fH)H[UHS1H M2fW /D$@HH $ $Yf.L$wHH[]@ATQUHSHPf.D$@D$E1 $N $D$8Y 1=1-C1X 0f(YL$\ 1\0f(=0\50^t$@0^X=0|$H\l$0f(Xl$ DHf(H\/$o$D$f(/fT.\D$ ^$XD$YXD$Xe0$H,f.V0d$r|$0f.H]-50f.v f.Ef(\$(D$D$H\$(t$@Y$D$^XL$HCH*X $\H*YT$8\T$f./AEĄt!fW\f.HPH[]A\f.x-D„uT$ $^ $T$f(f..sf.>.z u 1KSSf(HHf. .,f.vP\HL$f(H$eL$f(Qf.zaXYX$H[fDY -H$f(H$H*HH[XadH[Pf(\$\$fSf(HH $f(\$P$$Hf(f(Y$\$$HY[^f(D+SH\^f(H[Jf.SHH3HD$%L$H[^f(ÐSHHD$L$H$Y ?,f(L$L$f(Qf.zQf.z.$H[Y^D$f(H\$f(f(T$-T$f(뷐SHH0D$(,L$f.2,t$f.r=*|$^X|$ A@f(T$ $^f($XL* $fWT$\f.s{HY3,L$ Hf(YXXT$^\T$YL$ $ $+fW\T$Y\f.HHT$ZT$$f( *f. $wWT$( +X})fT$Xl+fW\[+$f.H0[fD x+fWf|$2+YY=)|$XQf.L$Xf(XQf.|$\f(X^f(YXX|$^|$ ` *H0[fWfD#X\g(Yo*H0[f(a $ $f(iHD$^D$\(HfHD$ '^L$Hf.HD$ *fW 'f(^L$H\f(f.HL$$ \(f.w6 .)\\f(YD$$H\f(DXYD$X$HDf.H $D$&\f()fW $\$HY\f(f.HL$$<&\^GL$YX$HHWH@f.HD$ )&\f(Y&Qf.zD$HYFf(f(Sf(Hf(H XL$T$^d$]T$L$f(Y 'YYf(YYXQf.z^\L$HT$YXL$ T$L$f(f(X^f.s Y^f(H f([f(T$\$`T$\$|Df.Uf(HSH(\=$&f(<$f(ؐH\$bHD$T-$%$^,$D$\d$f(f(H,5d$\$ $H*^T$XOT$L$\$Yf(^\$Yf(\ $^f.;H2H(H[]H#D$\$pL$$f.f(vfDYHXf.wHfH$"j#\f(-U#D$\$f(L$^f(HH,@f.$sf.AWIAVIAUIATIUHH)SHHH9H*|$fWLl$f.f(wDfWf.vEL $AH+ $H*f(^XH,H $H*\uD$\H,I)M9ILH[]A\A]A^A_ÐH*|$WL$AWIAVIAUATUHSHHhH9HNIH9H*MI*LMI)"I9HT$XLLOHt$PHL$HL)H*IV|$(^f(I*|$@YXT$H*Yp!\YH*Y^X K"Qf.HEMl$T$f(H*I*HD$0IFY=G#X=G#YH*|$ ^'L,T$IVH*ºf."Dф7fW|$f(HT$0L)H*ºf.x"DфfWM)XȺI*f.M"DфBfWMoXM)K.H*f."D„fWXL9L$8  yXL$@YG"L$1D$X L$]L$(HHD$z\ \$YD$ ^D$XD$f.wf.D$(sL,IFH*f.Y!EńfWHD$0L)H*f.2!Eń\fWMXM)IOH*f.!Ë́fWK .XH*f. Ë́fWXt$!l$8\\Y\ f.sBf(\Yf.f(l$@iXL$@f.DHD$PH9D$XMML)L9d$HLOHh[]A\A]A^LA_@f.H@D΄AL$@.L$@/f.@D΄L$@L$@f.DL$@L$@|@f.D7f(-XL$(f.DфvfWf(T$T$f.UD„T$8L$6L$T$8f.DфT$8L$L$T$8f.DфET$8L$L$T$8'f(f(^H  EDf.f(H8f(T$(\L$ \D$f(l$d$^4$4$d$f.l$\$L$ T$(s3 \f(\YYQf.z=\H8f(YYQf.z XH8f($$f($$뮐f.SHH D$\f(=D$Hf.D$H$YD$ 3 $\f(Yf.r\f(T$T$$f( $^XH,HjH [f.sHHat leastat mostexactlylongan integer is requirednumpy/random/mtrand/mtrand.c%s (%s:%d)Missing type objectnumpy%s.%s is not a type object%s() keywords must be stringscannot import name %.230sname '%s' is not definedmtrand.pyxmtrand.RandomState.__reduce__mtrand.RandomState.randnmtrand.RandomState.randmtrand.RandomState.__init__mtrand._shape_from_sizemtrand.RandomState.bytesmtrand.disc0_arraymtrand.RandomState.tomaxintmtrand.discd_array_scmtrand.cont2_array_scmtrand.RandomState.get_statemtrand.cont0_arraystandard_cauchystandard_exponentialstandard_normalrandom_samplemtrand.cont1_array_scmtrand.cont3_array_scmtrand.RandomState.set_statemtrand._rand_int8mtrand._rand_uint64mtrand._rand_uint16random_integersmtrand._rand_uint32mtrand._rand_boolmtrand._rand_int32mtrand._rand_int64mtrand._rand_int16mtrand._rand_uint8mtrand.cont1_arraymtrand.RandomState.rayleighmtrand.RandomState.powermtrand.RandomState.weibullmtrand.RandomState.paretomtrand.RandomState.standard_tmtrand.RandomState.chisquarestandard_gammamultinomialmtrand.discd_arraymtrand.RandomState.logseriesmtrand.RandomState.geometricmtrand.RandomState.zipfmtrand.RandomState.poissonmtrand.RandomState.seedmultivariate_normalmtrand.RandomState.randintmtrand.discnmN_arrayhypergeometricmtrand.discnmN_array_scfloat divisionmtrand.RandomState.dirichletmtrand.cont3_arraymtrand.RandomState.triangularnoncentral_fmtrand.cont2_arraymtrand.RandomState.waldmtrand.RandomState.lognormalmtrand.RandomState.logisticmtrand.RandomState.gumbelmtrand.RandomState.laplacemtrand.RandomState.vonmisesnoncentral_chisquaremtrand.RandomState.fmtrand.RandomState.gammamtrand.RandomState.betamtrand.RandomState.normalmtrand.RandomState.uniformmtrand.RandomState.shufflemtrand.discnp_array_scmtrand.discnp_arraymtrand.RandomState.binomialnegative_binomialmtrand.discdd_array_scmtrand.discdd_arraymtrand.RandomState.choiceassignmentnumpy.pxd%d.%d%s__builtin____builtins__214748364842949672954294967296-922337203685477580818446744073709551616__name__ndarrayflatiterbroadcast__pyx_vtable__numpy.core.multiarray_ARRAY_API_ARRAY_API not found_ARRAY_API is NULL pointerinit mtranddtypemtrand.RandomState__getstate____setstate__permutation%s() takes %s %zd positional argument%s (%zd given)need more than %zd value%s to unpack__%s__ returned non-%s (type %.200s)can't convert negative value to unsigned PY_LONG_LONGcan't convert negative value to unsigned longCannot convert %.200s to %.200stoo many values to unpack (expected %zd)%s.%s size changed, may indicate binary incompatibility%s.%s has the wrong size, try recompiling%s() got an unexpected keyword argument '%s'%s() got multiple values for keyword argument '%s'raise: arg 3 must be a traceback or Noneinstance exception may not have a separate valueraise: exception class must be a subclass of BaseExceptionmtrand.RandomState.__getstate__mtrand.RandomState.__setstate__mtrand.RandomState.permutationmtrand.RandomState.standard_cauchymtrand.RandomState.standard_exponentialmtrand.RandomState.standard_normalmtrand.RandomState.random_sample'%.200s' object is unsliceablevalue too large to convert to intcan't convert negative value to unsigned charvalue too large to convert to unsigned charvalue too large to convert to signed charcan't convert negative value to unsigned shortvalue too large to convert to unsigned shortmtrand.RandomState.random_integerscan't convert negative value to unsigned intvalue too large to convert to unsigned intvalue too large to convert to signed intvalue too large to convert to signed shortmtrand.RandomState.standard_gammamtrand.RandomState.exponentialmtrand.RandomState.multinomialmtrand.RandomState.multivariate_normalmtrand.RandomState.hypergeometricmtrand.RandomState.noncentral_fmtrand.RandomState.noncentral_chisquaremtrand.RandomState.negative_binomial'%.200s' object does not support slice %scompiletime version %s of module '%.100s' does not match runtime version %snumpy.core.multiarray failed to import_ARRAY_API is not PyCObject objectmodule compiled against ABI version 0x%x but this version of numpy is 0x%xmodule compiled against API version 0x%x but this version of numpy is 0x%xFATAL: module compiled as unknown endianFATAL: module compiled as little endian, but detected different endianness at runtime RandomState(seed=None) Container for the Mersenne Twister pseudo-random number generator. `RandomState` exposes a number of methods for generating random numbers drawn from a variety of probability distributions. In addition to the distribution-specific arguments, each method takes a keyword argument `size` that defaults to ``None``. If `size` is ``None``, then a single value is generated and returned. If `size` is an integer, then a 1-D array filled with generated values is returned. If `size` is a tuple, then an array with that shape is filled and returned. *Compatibility Guarantee* A fixed seed and a fixed series of calls to 'RandomState' methods using the same parameters will always produce the same results up to roundoff error except when the values were incorrect. Incorrect values will be fixed and the NumPy version in which the fix was made will be noted in the relevant docstring. Extension of existing parameter ranges and the addition of new parameters is allowed as long the previous behavior remains unchanged. Parameters ---------- seed : {None, int, array_like}, optional Random seed initializing the pseudo-random number generator. Can be an integer, an array (or other sequence) of integers of any length, or ``None`` (the default). If `seed` is ``None``, then `RandomState` will try to read data from ``/dev/urandom`` (or the Windows analogue) if available or seed from the clock otherwise. Notes ----- The Python stdlib module "random" also contains a Mersenne Twister pseudo-random number generator with a number of methods that are similar to the ones available in `RandomState`. `RandomState`, besides being NumPy-aware, has the advantage that it provides a much larger number of probability distributions to choose from. _____hh}hthkhqqqqqdžߏ_VMD;{ri`Wwne\Szqh_''''y':::::??rb/dev/random/dev/urandomno errorrandom device unvavailableA<UUUUUU?llfJ?88C$+K?<ٰj_AAz?SˆB8?5gG5gG@dg??UUUUUU?"@m{?(\@ffffff@.@4@x&?@?UUUUUU?a@X@`@|@@MA$@>@= ףp=@n?[ m?h|?5?333333 @r?$~?@B>٬ @r鷯?Q?Q?9v?:0yE>-DT! @h㈵>@-DT!@3?r?q?0@;4P0a(^< H p `  H xppPh0 `08PP @ ( 0H x @ `8 8 0 @ (xp``&P(*0+p-p4p;PV_i@`r0~0А Pp ` pP`@`"0/@<D Spv`P@)<@PM[`e0 ox  p  `@ P  . @08!@W! !("p@"X""#`h# #$в0$H$x$$$(%p% %`%%к(&@&0X&Pp&&&&'('H'''(((``(x(p(((0)8)pp))) )))@*` *@*h** ***(+x+ ++zRx $@ FJ w?;*3$"4DEBDD a GBL AAB,|iADD L CAG 4AAD H AAH O AAG 4PAAD H AAH O AAG v4=GuL8lD X D l1A,PAD _ AA R AE .<BED D(DP (A ABBI ,,hYAD h AB X AG 4\BBA A(A0(A ABBD`BFB B(A0A8D`8A0A(B BBB$AG0 AG &dWBBB E(D0D8D 8A0A(B BBBF O 8A0A(B BBBG AD A~oAD hAͽoAD hA,[BDG D0E AAB$GYAGD0GDA,<xpBDA cCBlII XE bDBEL A(A0G 0A(A BBBH <BEE A(A0DP0A(A BBBL;BEB B(A0E8Gz8A0A(B BBB,dAADP AAG ,0|ADD ^ DAA ,vADD h DAA )A\ C D,AAD0b AAA ,D5BKA  ABF LtBBB E(A0A8DP 8D0A(B BBBH <zBBD A(D0 (D ABBF \ }BBB A(D0G@ 0A(A BBBI  0A(A BBBH \d@BBB A(D0G@ 0A(A BBBG  0A(A BBBH <BBA D(GP (A ABBE L`BBB B(A0D8G 8A0A(B BBBK LTBBE B(A0A8D` 8A0A(B BBBC L`BBB B(A0D8G 8A0A(B BBBJ LBBB E(D0D8Dp# 8A0A(B BBBD DD ptBEE A(A0D@( 0A(A BBBF L BBE E(D0A8Dy 8D0A(B BBBC L HI BBB E(A0A8DP 8A0A(B BBBH L, HBBE E(D0A8Dpe 8D0A(B BBBG <| BBA D(G@ (A ABBJ < XBBA D(G@ (A ABBJ < BBA D(G@ (A ABBJ << XBBA D(G@ (A ABBJ L| BBE E(D0A8Dpn 8D0A(B BBBF L h#BBE E(D0A8D 8D0A(B BBBJ L *BEB B(A0D8D 8A0A(B BBBD Ll DY BBB B(A0D8Gr 8A0A(B BBBH L M BBB B(A0D8GV 8A0A(B BBBD L VJ BBB B(A0D8Gv 8A0A(B BBBD L\ _ BBB B(A0D8G 8A0A(B BBBE L HkJ BBB B(A0D8Go 8A0A(B BBBC L HtJ BBB B(A0D8Gq 8A0A(B BBBA LLH}@ BBB B(A0D8Gp 8A0A(B BBBB L8BBB B(A0D8GW 8A0A(B BBBK LȎY BBB B(A0D8Gw 8A0A(B BBBK L<ؗJ BBB B(A0D8Gq 8A0A(B BBBA LؠzBBB B(D0D8Dd 8D0A(B BBBK L_BBB E(A0A8Gp 8A0A(B BBBH L,PBBB E(A0D8Gp  8A0A(B BBBG L|PBBB E(A0D8Gp  8A0A(B BBBG LPBBB E(A0D8Gp  8A0A(B BBBG LPBBB E(A0D8Gp  8A0A(B BBBG LlPBBB E(A0D8Gp  8A0A(B BBBG LPBBB E(A0D8Gp  8A0A(B BBBG L _BBB E(A0A8Gp 8A0A(B BBBH L\( BBB E(A0D8J4 8A0A(B BBBA LxZBBE B(D0D8G 8D0A(B BBBD L BBB E(A0D8Gp 8A0A(B BBBG LL BBB E(A0D8Gp 8A0A(B BBBG Lh#PBBB E(A0D8Gp( 8A0A(B BBBG Lh+BBB B(D0D8D 8A0A(B BBBH L<:}"BBB B(A0A8G 8A0A(B BBBJ L8\1BBB B(D0D8Jx 8A0A(B BBBD L!BBE B(A0D8J 8A0A(B BBBA L,h BBE E(A0D8G4 8D0A(B BBBE L|V BBB B(A0D8J 8A0A(B BBBD LBBB B(D0D8J 8A0A(B BBBA Lh@BBE B(D0A8D 8D0A(B BBBH LlX9BBB B(D0A8Gc 8A0A(B BBBG LH BBB B(D0A8G[ 8A0A(B BBBG L  [BBB E(D0A8D 8A0A(B BBBG L\0FBBB B(A0D8G; 8A0A(B BBBG L> BBB B(A0A8G 8A0A(B BBBJ L(H BBB B(A0A8G 8A0A(B BBBJ LLQ BBB B(A0A8G 8A0A(B BBBJ L[ BBB B(A0A8G 8A0A(B BBBJ Lxdw BBB B(A0D8G 8A0A(B BBBG L<m BBB B(A0D8G; 8A0A(B BBBG LX{ BBB B(A0D8G; 8A0A(B BBBG L BBB B(A0D8G 8A0A(B BBBG L, BBB B(A0D8G+ 8A0A(B BBBG L| BBB B(A0A8G 8A0A(B BBBJ L8kBBE B(A0A8G8 8A0A(B BBBJ LXb1BEB H(A0A8D 8A0A(B BBBA Llx'BBB B(D0D8GM 8D0A(B BBBG D( BBB D(D0GP 0A(A BBBA L'BBB B(A0D8J 8A0A(B BBBD LT5ַBBB B(A0D8JN 8A0A(B BBBA L@BYB B(A0A8G 8A0A(B BBBD _ (<L$P$BBB B(A0D8D`} 8A0A(B BBBH \t0BEE B(A0A8FPj 8A0A(B BBBA Y8A0A(B BBB\BBB B(D0A8D@` 8A0A(B BBBA 8A0A(B BBB\4BBB B(D0A8D@_ 8A0A(B BBBA s8A0A(B BBB<P~GDE QABL p$-AAG WHDDI,BFA h ABD $DFAAG zAALl8BIE D(A0V (A BBBF c(A BBB4vBMD u EEF UAGDBIB A(A0GP 0A(A BBBA ,< h9BDD G0c AAB,l x6AG0 AD a AA $ >yVA, +AED@ EAB  (D c !/Df$!ДD U>> np.random.permutation(10) array([1, 7, 4, 3, 0, 9, 2, 5, 8, 6]) >>> np.random.permutation([1, 4, 9, 12, 15]) array([15, 1, 9, 4, 12]) >>> arr = np.arange(9).reshape((3, 3)) >>> np.random.permutation(arr) array([[6, 7, 8], [0, 1, 2], [3, 4, 5]]) shuffle(x) Modify a sequence in-place by shuffling its contents. Parameters ---------- x : array_like The array or list to be shuffled. Returns ------- None Examples -------- >>> arr = np.arange(10) >>> np.random.shuffle(arr) >>> arr [1 7 5 2 9 4 3 6 0 8] This function only shuffles the array along the first index of a multi-dimensional array: >>> arr = np.arange(9).reshape((3, 3)) >>> np.random.shuffle(arr) >>> arr array([[3, 4, 5], [6, 7, 8], [0, 1, 2]]) dirichlet(alpha, size=None) Draw samples from the Dirichlet distribution. Draw `size` samples of dimension k from a Dirichlet distribution. A Dirichlet-distributed random variable can be seen as a multivariate generalization of a Beta distribution. Dirichlet pdf is the conjugate prior of a multinomial in Bayesian inference. Parameters ---------- alpha : array Parameter of the distribution (k dimension for sample of dimension k). size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray, The drawn samples, of shape (size, alpha.ndim). Notes ----- .. math:: X \approx \prod_{i=1}^{k}{x^{\alpha_i-1}_i} Uses the following property for computation: for each dimension, draw a random sample y_i from a standard gamma generator of shape `alpha_i`, then :math:`X = \frac{1}{\sum_{i=1}^k{y_i}} (y_1, \ldots, y_n)` is Dirichlet distributed. References ---------- .. [1] David McKay, "Information Theory, Inference and Learning Algorithms," chapter 23, http://www.inference.phy.cam.ac.uk/mackay/ .. [2] Wikipedia, "Dirichlet distribution", http://en.wikipedia.org/wiki/Dirichlet_distribution Examples -------- Taking an example cited in Wikipedia, this distribution can be used if one wanted to cut strings (each of initial length 1.0) into K pieces with different lengths, where each piece had, on average, a designated average length, but allowing some variation in the relative sizes of the pieces. >>> s = np.random.dirichlet((10, 5, 3), 20).transpose() >>> plt.barh(range(20), s[0]) >>> plt.barh(range(20), s[1], left=s[0], color='g') >>> plt.barh(range(20), s[2], left=s[0]+s[1], color='r') >>> plt.title("Lengths of Strings") multinomial(n, pvals, size=None) Draw samples from a multinomial distribution. The multinomial distribution is a multivariate generalisation of the binomial distribution. Take an experiment with one of ``p`` possible outcomes. An example of such an experiment is throwing a dice, where the outcome can be 1 through 6. Each sample drawn from the distribution represents `n` such experiments. Its values, ``X_i = [X_0, X_1, ..., X_p]``, represent the number of times the outcome was ``i``. Parameters ---------- n : int Number of experiments. pvals : sequence of floats, length p Probabilities of each of the ``p`` different outcomes. These should sum to 1 (however, the last element is always assumed to account for the remaining probability, as long as ``sum(pvals[:-1]) <= 1)``. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : ndarray The drawn samples, of shape *size*, if that was provided. If not, the shape is ``(N,)``. In other words, each entry ``out[i,j,...,:]`` is an N-dimensional value drawn from the distribution. Examples -------- Throw a dice 20 times: >>> np.random.multinomial(20, [1/6.]*6, size=1) array([[4, 1, 7, 5, 2, 1]]) It landed 4 times on 1, once on 2, etc. Now, throw the dice 20 times, and 20 times again: >>> np.random.multinomial(20, [1/6.]*6, size=2) array([[3, 4, 3, 3, 4, 3], [2, 4, 3, 4, 0, 7]]) For the first run, we threw 3 times 1, 4 times 2, etc. For the second, we threw 2 times 1, 4 times 2, etc. A loaded die is more likely to land on number 6: >>> np.random.multinomial(100, [1/7.]*5 + [2/7.]) array([11, 16, 14, 17, 16, 26]) The probability inputs should be normalized. As an implementation detail, the value of the last entry is ignored and assumed to take up any leftover probability mass, but this should not be relied on. A biased coin which has twice as much weight on one side as on the other should be sampled like so: >>> np.random.multinomial(100, [1.0 / 3, 2.0 / 3]) # RIGHT array([38, 62]) not like: >>> np.random.multinomial(100, [1.0, 2.0]) # WRONG array([100, 0]) multivariate_normal(mean, cov[, size]) Draw random samples from a multivariate normal distribution. The multivariate normal, multinormal or Gaussian distribution is a generalization of the one-dimensional normal distribution to higher dimensions. Such a distribution is specified by its mean and covariance matrix. These parameters are analogous to the mean (average or "center") and variance (standard deviation, or "width," squared) of the one-dimensional normal distribution. Parameters ---------- mean : 1-D array_like, of length N Mean of the N-dimensional distribution. cov : 2-D array_like, of shape (N, N) Covariance matrix of the distribution. It must be symmetric and positive-semidefinite for proper sampling. size : int or tuple of ints, optional Given a shape of, for example, ``(m,n,k)``, ``m*n*k`` samples are generated, and packed in an `m`-by-`n`-by-`k` arrangement. Because each sample is `N`-dimensional, the output shape is ``(m,n,k,N)``. If no shape is specified, a single (`N`-D) sample is returned. Returns ------- out : ndarray The drawn samples, of shape *size*, if that was provided. If not, the shape is ``(N,)``. In other words, each entry ``out[i,j,...,:]`` is an N-dimensional value drawn from the distribution. Notes ----- The mean is a coordinate in N-dimensional space, which represents the location where samples are most likely to be generated. This is analogous to the peak of the bell curve for the one-dimensional or univariate normal distribution. Covariance indicates the level to which two variables vary together. From the multivariate normal distribution, we draw N-dimensional samples, :math:`X = [x_1, x_2, ... x_N]`. The covariance matrix element :math:`C_{ij}` is the covariance of :math:`x_i` and :math:`x_j`. The element :math:`C_{ii}` is the variance of :math:`x_i` (i.e. its "spread"). Instead of specifying the full covariance matrix, popular approximations include: - Spherical covariance (*cov* is a multiple of the identity matrix) - Diagonal covariance (*cov* has non-negative elements, and only on the diagonal) This geometrical property can be seen in two dimensions by plotting generated data-points: >>> mean = [0, 0] >>> cov = [[1, 0], [0, 100]] # diagonal covariance Diagonal covariance means that points are oriented along x or y-axis: >>> import matplotlib.pyplot as plt >>> x, y = np.random.multivariate_normal(mean, cov, 5000).T >>> plt.plot(x, y, 'x') >>> plt.axis('equal') >>> plt.show() Note that the covariance matrix must be positive semidefinite (a.k.a. nonnegative-definite). Otherwise, the behavior of this method is undefined and backwards compatibility is not guaranteed. References ---------- .. [1] Papoulis, A., "Probability, Random Variables, and Stochastic Processes," 3rd ed., New York: McGraw-Hill, 1991. .. [2] Duda, R. O., Hart, P. E., and Stork, D. G., "Pattern Classification," 2nd ed., New York: Wiley, 2001. Examples -------- >>> mean = (1, 2) >>> cov = [[1, 0], [0, 1]] >>> x = np.random.multivariate_normal(mean, cov, (3, 3)) >>> x.shape (3, 3, 2) The following is probably true, given that 0.6 is roughly twice the standard deviation: >>> list((x[0,0,:] - mean) < 0.6) [True, True] logseries(p, size=None) Draw samples from a logarithmic series distribution. Samples are drawn from a log series distribution with specified shape parameter, 0 < ``p`` < 1. Parameters ---------- loc : float scale : float > 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar where the values are all integers in [0, n]. See Also -------- scipy.stats.distributions.logser : probability density function, distribution or cumulative density function, etc. Notes ----- The probability density for the Log Series distribution is .. math:: P(k) = \frac{-p^k}{k \ln(1-p)}, where p = probability. The log series distribution is frequently used to represent species richness and occurrence, first proposed by Fisher, Corbet, and Williams in 1943 [2]. It may also be used to model the numbers of occupants seen in cars [3]. References ---------- .. [1] Buzas, Martin A.; Culver, Stephen J., Understanding regional species diversity through the log series distribution of occurrences: BIODIVERSITY RESEARCH Diversity & Distributions, Volume 5, Number 5, September 1999 , pp. 187-195(9). .. [2] Fisher, R.A,, A.S. Corbet, and C.B. Williams. 1943. The relation between the number of species and the number of individuals in a random sample of an animal population. Journal of Animal Ecology, 12:42-58. .. [3] D. J. Hand, F. Daly, D. Lunn, E. Ostrowski, A Handbook of Small Data Sets, CRC Press, 1994. .. [4] Wikipedia, "Logarithmic-distribution", http://en.wikipedia.org/wiki/Logarithmic-distribution Examples -------- Draw samples from the distribution: >>> a = .6 >>> s = np.random.logseries(a, 10000) >>> count, bins, ignored = plt.hist(s) # plot against distribution >>> def logseries(k, p): ... return -p**k/(k*log(1-p)) >>> plt.plot(bins, logseries(bins, a)*count.max()/ logseries(bins, a).max(), 'r') >>> plt.show() hypergeometric(ngood, nbad, nsample, size=None) Draw samples from a Hypergeometric distribution. Samples are drawn from a hypergeometric distribution with specified parameters, ngood (ways to make a good selection), nbad (ways to make a bad selection), and nsample = number of items sampled, which is less than or equal to the sum ngood + nbad. Parameters ---------- ngood : int or array_like Number of ways to make a good selection. Must be nonnegative. nbad : int or array_like Number of ways to make a bad selection. Must be nonnegative. nsample : int or array_like Number of items sampled. Must be at least 1 and at most ``ngood + nbad``. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar The values are all integers in [0, n]. See Also -------- scipy.stats.distributions.hypergeom : probability density function, distribution or cumulative density function, etc. Notes ----- The probability density for the Hypergeometric distribution is .. math:: P(x) = \frac{\binom{m}{n}\binom{N-m}{n-x}}{\binom{N}{n}}, where :math:`0 \le x \le m` and :math:`n+m-N \le x \le n` for P(x) the probability of x successes, n = ngood, m = nbad, and N = number of samples. Consider an urn with black and white marbles in it, ngood of them black and nbad are white. If you draw nsample balls without replacement, then the hypergeometric distribution describes the distribution of black balls in the drawn sample. Note that this distribution is very similar to the binomial distribution, except that in this case, samples are drawn without replacement, whereas in the Binomial case samples are drawn with replacement (or the sample space is infinite). As the sample space becomes large, this distribution approaches the binomial. References ---------- .. [1] Lentner, Marvin, "Elementary Applied Statistics", Bogden and Quigley, 1972. .. [2] Weisstein, Eric W. "Hypergeometric Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/HypergeometricDistribution.html .. [3] Wikipedia, "Hypergeometric-distribution", http://en.wikipedia.org/wiki/Hypergeometric_distribution Examples -------- Draw samples from the distribution: >>> ngood, nbad, nsamp = 100, 2, 10 # number of good, number of bad, and number of samples >>> s = np.random.hypergeometric(ngood, nbad, nsamp, 1000) >>> hist(s) # note that it is very unlikely to grab both bad items Suppose you have an urn with 15 white and 15 black marbles. If you pull 15 marbles at random, how likely is it that 12 or more of them are one color? >>> s = np.random.hypergeometric(15, 15, 15, 100000) >>> sum(s>=12)/100000. + sum(s<=3)/100000. # answer = 0.003 ... pretty unlikely! geometric(p, size=None) Draw samples from the geometric distribution. Bernoulli trials are experiments with one of two outcomes: success or failure (an example of such an experiment is flipping a coin). The geometric distribution models the number of trials that must be run in order to achieve success. It is therefore supported on the positive integers, ``k = 1, 2, ...``. The probability mass function of the geometric distribution is .. math:: f(k) = (1 - p)^{k - 1} p where `p` is the probability of success of an individual trial. Parameters ---------- p : float The probability of success of an individual trial. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : ndarray Samples from the geometric distribution, shaped according to `size`. Examples -------- Draw ten thousand values from the geometric distribution, with the probability of an individual success equal to 0.35: >>> z = np.random.geometric(p=0.35, size=10000) How many trials succeeded after a single run? >>> (z == 1).sum() / 10000. 0.34889999999999999 #random zipf(a, size=None) Draw samples from a Zipf distribution. Samples are drawn from a Zipf distribution with specified parameter `a` > 1. The Zipf distribution (also known as the zeta distribution) is a continuous probability distribution that satisfies Zipf's law: the frequency of an item is inversely proportional to its rank in a frequency table. Parameters ---------- a : float > 1 Distribution parameter. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : scalar or ndarray The returned samples are greater than or equal to one. See Also -------- scipy.stats.distributions.zipf : probability density function, distribution, or cumulative density function, etc. Notes ----- The probability density for the Zipf distribution is .. math:: p(x) = \frac{x^{-a}}{\zeta(a)}, where :math:`\zeta` is the Riemann Zeta function. It is named for the American linguist George Kingsley Zipf, who noted that the frequency of any word in a sample of a language is inversely proportional to its rank in the frequency table. References ---------- .. [1] Zipf, G. K., "Selected Studies of the Principle of Relative Frequency in Language," Cambridge, MA: Harvard Univ. Press, 1932. Examples -------- Draw samples from the distribution: >>> a = 2. # parameter >>> s = np.random.zipf(a, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> import scipy.special as sps Truncate s values at 50 so plot is interesting >>> count, bins, ignored = plt.hist(s[s<50], 50, normed=True) >>> x = np.arange(1., 50.) >>> y = x**(-a)/sps.zetac(a) >>> plt.plot(x, y/max(y), linewidth=2, color='r') >>> plt.show() poisson(lam=1.0, size=None) Draw samples from a Poisson distribution. The Poisson distribution is the limit of the binomial distribution for large N. Parameters ---------- lam : float or sequence of float Expectation of interval, should be >= 0. A sequence of expectation intervals must be broadcastable over the requested size. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar The drawn samples, of shape *size*, if it was provided. Notes ----- The Poisson distribution .. math:: f(k; \lambda)=\frac{\lambda^k e^{-\lambda}}{k!} For events with an expected separation :math:`\lambda` the Poisson distribution :math:`f(k; \lambda)` describes the probability of :math:`k` events occurring within the observed interval :math:`\lambda`. Because the output is limited to the range of the C long type, a ValueError is raised when `lam` is within 10 sigma of the maximum representable value. References ---------- .. [1] Weisstein, Eric W. "Poisson Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/PoissonDistribution.html .. [2] Wikipedia, "Poisson distribution", http://en.wikipedia.org/wiki/Poisson_distribution Examples -------- Draw samples from the distribution: >>> import numpy as np >>> s = np.random.poisson(5, 10000) Display histogram of the sample: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 14, normed=True) >>> plt.show() Draw each 100 values for lambda 100 and 500: >>> s = np.random.poisson(lam=(100., 500.), size=(100, 2)) negative_binomial(n, p, size=None) Draw samples from a negative binomial distribution. Samples are drawn from a negative binomial distribution with specified parameters, `n` trials and `p` probability of success where `n` is an integer > 0 and `p` is in the interval [0, 1]. Parameters ---------- n : int Parameter, > 0. p : float Parameter, >= 0 and <=1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : int or ndarray of ints Drawn samples. Notes ----- The probability density for the negative binomial distribution is .. math:: P(N;n,p) = \binom{N+n-1}{n-1}p^{n}(1-p)^{N}, where :math:`n-1` is the number of successes, :math:`p` is the probability of success, and :math:`N+n-1` is the number of trials. The negative binomial distribution gives the probability of n-1 successes and N failures in N+n-1 trials, and success on the (N+n)th trial. If one throws a die repeatedly until the third time a "1" appears, then the probability distribution of the number of non-"1"s that appear before the third "1" is a negative binomial distribution. References ---------- .. [1] Weisstein, Eric W. "Negative Binomial Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/NegativeBinomialDistribution.html .. [2] Wikipedia, "Negative binomial distribution", http://en.wikipedia.org/wiki/Negative_binomial_distribution Examples -------- Draw samples from the distribution: A real world example. A company drills wild-cat oil exploration wells, each with an estimated probability of success of 0.1. What is the probability of having one success for each successive well, that is what is the probability of a single success after drilling 5 wells, after 6 wells, etc.? >>> s = np.random.negative_binomial(1, 0.1, 100000) >>> for i in range(1, 11): ... probability = sum(s>> n, p = 10, .5 # number of trials, probability of each trial >>> s = np.random.binomial(n, p, 1000) # result of flipping a coin 10 times, tested 1000 times. A real world example. A company drills 9 wild-cat oil exploration wells, each with an estimated probability of success of 0.1. All nine wells fail. What is the probability of that happening? Let's do 20,000 trials of the model, and count the number that generate zero positive results. >>> sum(np.random.binomial(9, 0.1, 20000) == 0)/20000. # answer = 0.38885, or 38%. triangular(left, mode, right, size=None) Draw samples from the triangular distribution. The triangular distribution is a continuous probability distribution with lower limit left, peak at mode, and upper limit right. Unlike the other distributions, these parameters directly define the shape of the pdf. Parameters ---------- left : scalar Lower limit. mode : scalar The value where the peak of the distribution occurs. The value should fulfill the condition ``left <= mode <= right``. right : scalar Upper limit, should be larger than `left`. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar The returned samples all lie in the interval [left, right]. Notes ----- The probability density function for the triangular distribution is .. math:: P(x;l, m, r) = \begin{cases} \frac{2(x-l)}{(r-l)(m-l)}& \text{for $l \leq x \leq m$},\\ \frac{2(r-x)}{(r-l)(r-m)}& \text{for $m \leq x \leq r$},\\ 0& \text{otherwise}. \end{cases} The triangular distribution is often used in ill-defined problems where the underlying distribution is not known, but some knowledge of the limits and mode exists. Often it is used in simulations. References ---------- .. [1] Wikipedia, "Triangular distribution" http://en.wikipedia.org/wiki/Triangular_distribution Examples -------- Draw values from the distribution and plot the histogram: >>> import matplotlib.pyplot as plt >>> h = plt.hist(np.random.triangular(-3, 0, 8, 100000), bins=200, ... normed=True) >>> plt.show() wald(mean, scale, size=None) Draw samples from a Wald, or inverse Gaussian, distribution. As the scale approaches infinity, the distribution becomes more like a Gaussian. Some references claim that the Wald is an inverse Gaussian with mean equal to 1, but this is by no means universal. The inverse Gaussian distribution was first studied in relationship to Brownian motion. In 1956 M.C.K. Tweedie used the name inverse Gaussian because there is an inverse relationship between the time to cover a unit distance and distance covered in unit time. Parameters ---------- mean : scalar Distribution mean, should be > 0. scale : scalar Scale parameter, should be >= 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar Drawn sample, all greater than zero. Notes ----- The probability density function for the Wald distribution is .. math:: P(x;mean,scale) = \sqrt{\frac{scale}{2\pi x^3}}e^ \frac{-scale(x-mean)^2}{2\cdotp mean^2x} As noted above the inverse Gaussian distribution first arise from attempts to model Brownian motion. It is also a competitor to the Weibull for use in reliability modeling and modeling stock returns and interest rate processes. References ---------- .. [1] Brighton Webs Ltd., Wald Distribution, http://www.brighton-webs.co.uk/distributions/wald.asp .. [2] Chhikara, Raj S., and Folks, J. Leroy, "The Inverse Gaussian Distribution: Theory : Methodology, and Applications", CRC Press, 1988. .. [3] Wikipedia, "Wald distribution" http://en.wikipedia.org/wiki/Wald_distribution Examples -------- Draw values from the distribution and plot the histogram: >>> import matplotlib.pyplot as plt >>> h = plt.hist(np.random.wald(3, 2, 100000), bins=200, normed=True) >>> plt.show() rayleigh(scale=1.0, size=None) Draw samples from a Rayleigh distribution. The :math:`\chi` and Weibull distributions are generalizations of the Rayleigh. Parameters ---------- scale : scalar Scale, also equals the mode. Should be >= 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Notes ----- The probability density function for the Rayleigh distribution is .. math:: P(x;scale) = \frac{x}{scale^2}e^{\frac{-x^2}{2 \cdotp scale^2}} The Rayleigh distribution would arise, for example, if the East and North components of the wind velocity had identical zero-mean Gaussian distributions. Then the wind speed would have a Rayleigh distribution. References ---------- .. [1] Brighton Webs Ltd., "Rayleigh Distribution," http://www.brighton-webs.co.uk/distributions/rayleigh.asp .. [2] Wikipedia, "Rayleigh distribution" http://en.wikipedia.org/wiki/Rayleigh_distribution Examples -------- Draw values from the distribution and plot the histogram >>> values = hist(np.random.rayleigh(3, 100000), bins=200, normed=True) Wave heights tend to follow a Rayleigh distribution. If the mean wave height is 1 meter, what fraction of waves are likely to be larger than 3 meters? >>> meanvalue = 1 >>> modevalue = np.sqrt(2 / np.pi) * meanvalue >>> s = np.random.rayleigh(modevalue, 1000000) The percentage of waves larger than 3 meters is: >>> 100.*sum(s>3)/1000000. 0.087300000000000003 lognormal(mean=0.0, sigma=1.0, size=None) Draw samples from a log-normal distribution. Draw samples from a log-normal distribution with specified mean, standard deviation, and array shape. Note that the mean and standard deviation are not the values for the distribution itself, but of the underlying normal distribution it is derived from. Parameters ---------- mean : float Mean value of the underlying normal distribution sigma : float, > 0. Standard deviation of the underlying normal distribution size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or float The desired samples. An array of the same shape as `size` if given, if `size` is None a float is returned. See Also -------- scipy.stats.lognorm : probability density function, distribution, cumulative density function, etc. Notes ----- A variable `x` has a log-normal distribution if `log(x)` is normally distributed. The probability density function for the log-normal distribution is: .. math:: p(x) = \frac{1}{\sigma x \sqrt{2\pi}} e^{(-\frac{(ln(x)-\mu)^2}{2\sigma^2})} where :math:`\mu` is the mean and :math:`\sigma` is the standard deviation of the normally distributed logarithm of the variable. A log-normal distribution results if a random variable is the *product* of a large number of independent, identically-distributed variables in the same way that a normal distribution results if the variable is the *sum* of a large number of independent, identically-distributed variables. References ---------- .. [1] Limpert, E., Stahel, W. A., and Abbt, M., "Log-normal Distributions across the Sciences: Keys and Clues," BioScience, Vol. 51, No. 5, May, 2001. http://stat.ethz.ch/~stahel/lognormal/bioscience.pdf .. [2] Reiss, R.D. and Thomas, M., "Statistical Analysis of Extreme Values," Basel: Birkhauser Verlag, 2001, pp. 31-32. Examples -------- Draw samples from the distribution: >>> mu, sigma = 3., 1. # mean and standard deviation >>> s = np.random.lognormal(mu, sigma, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 100, normed=True, align='mid') >>> x = np.linspace(min(bins), max(bins), 10000) >>> pdf = (np.exp(-(np.log(x) - mu)**2 / (2 * sigma**2)) ... / (x * sigma * np.sqrt(2 * np.pi))) >>> plt.plot(x, pdf, linewidth=2, color='r') >>> plt.axis('tight') >>> plt.show() Demonstrate that taking the products of random samples from a uniform distribution can be fit well by a log-normal probability density function. >>> # Generate a thousand samples: each is the product of 100 random >>> # values, drawn from a normal distribution. >>> b = [] >>> for i in range(1000): ... a = 10. + np.random.random(100) ... b.append(np.product(a)) >>> b = np.array(b) / np.min(b) # scale values to be positive >>> count, bins, ignored = plt.hist(b, 100, normed=True, align='mid') >>> sigma = np.std(np.log(b)) >>> mu = np.mean(np.log(b)) >>> x = np.linspace(min(bins), max(bins), 10000) >>> pdf = (np.exp(-(np.log(x) - mu)**2 / (2 * sigma**2)) ... / (x * sigma * np.sqrt(2 * np.pi))) >>> plt.plot(x, pdf, color='r', linewidth=2) >>> plt.show() logistic(loc=0.0, scale=1.0, size=None) Draw samples from a logistic distribution. Samples are drawn from a logistic distribution with specified parameters, loc (location or mean, also median), and scale (>0). Parameters ---------- loc : float scale : float > 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar where the values are all integers in [0, n]. See Also -------- scipy.stats.distributions.logistic : probability density function, distribution or cumulative density function, etc. Notes ----- The probability density for the Logistic distribution is .. math:: P(x) = P(x) = \frac{e^{-(x-\mu)/s}}{s(1+e^{-(x-\mu)/s})^2}, where :math:`\mu` = location and :math:`s` = scale. The Logistic distribution is used in Extreme Value problems where it can act as a mixture of Gumbel distributions, in Epidemiology, and by the World Chess Federation (FIDE) where it is used in the Elo ranking system, assuming the performance of each player is a logistically distributed random variable. References ---------- .. [1] Reiss, R.-D. and Thomas M. (2001), "Statistical Analysis of Extreme Values, from Insurance, Finance, Hydrology and Other Fields," Birkhauser Verlag, Basel, pp 132-133. .. [2] Weisstein, Eric W. "Logistic Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/LogisticDistribution.html .. [3] Wikipedia, "Logistic-distribution", http://en.wikipedia.org/wiki/Logistic_distribution Examples -------- Draw samples from the distribution: >>> loc, scale = 10, 1 >>> s = np.random.logistic(loc, scale, 10000) >>> count, bins, ignored = plt.hist(s, bins=50) # plot against distribution >>> def logist(x, loc, scale): ... return exp((loc-x)/scale)/(scale*(1+exp((loc-x)/scale))**2) >>> plt.plot(bins, logist(bins, loc, scale)*count.max()/\ ... logist(bins, loc, scale).max()) >>> plt.show() gumbel(loc=0.0, scale=1.0, size=None) Draw samples from a Gumbel distribution. Draw samples from a Gumbel distribution with specified location and scale. For more information on the Gumbel distribution, see Notes and References below. Parameters ---------- loc : float The location of the mode of the distribution. scale : float The scale parameter of the distribution. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar See Also -------- scipy.stats.gumbel_l scipy.stats.gumbel_r scipy.stats.genextreme weibull Notes ----- The Gumbel (or Smallest Extreme Value (SEV) or the Smallest Extreme Value Type I) distribution is one of a class of Generalized Extreme Value (GEV) distributions used in modeling extreme value problems. The Gumbel is a special case of the Extreme Value Type I distribution for maximums from distributions with "exponential-like" tails. The probability density for the Gumbel distribution is .. math:: p(x) = \frac{e^{-(x - \mu)/ \beta}}{\beta} e^{ -e^{-(x - \mu)/ \beta}}, where :math:`\mu` is the mode, a location parameter, and :math:`\beta` is the scale parameter. The Gumbel (named for German mathematician Emil Julius Gumbel) was used very early in the hydrology literature, for modeling the occurrence of flood events. It is also used for modeling maximum wind speed and rainfall rates. It is a "fat-tailed" distribution - the probability of an event in the tail of the distribution is larger than if one used a Gaussian, hence the surprisingly frequent occurrence of 100-year floods. Floods were initially modeled as a Gaussian process, which underestimated the frequency of extreme events. It is one of a class of extreme value distributions, the Generalized Extreme Value (GEV) distributions, which also includes the Weibull and Frechet. The function has a mean of :math:`\mu + 0.57721\beta` and a variance of :math:`\frac{\pi^2}{6}\beta^2`. References ---------- .. [1] Gumbel, E. J., "Statistics of Extremes," New York: Columbia University Press, 1958. .. [2] Reiss, R.-D. and Thomas, M., "Statistical Analysis of Extreme Values from Insurance, Finance, Hydrology and Other Fields," Basel: Birkhauser Verlag, 2001. Examples -------- Draw samples from the distribution: >>> mu, beta = 0, 0.1 # location and scale >>> s = np.random.gumbel(mu, beta, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 30, normed=True) >>> plt.plot(bins, (1/beta)*np.exp(-(bins - mu)/beta) ... * np.exp( -np.exp( -(bins - mu) /beta) ), ... linewidth=2, color='r') >>> plt.show() Show how an extreme value distribution can arise from a Gaussian process and compare to a Gaussian: >>> means = [] >>> maxima = [] >>> for i in range(0,1000) : ... a = np.random.normal(mu, beta, 1000) ... means.append(a.mean()) ... maxima.append(a.max()) >>> count, bins, ignored = plt.hist(maxima, 30, normed=True) >>> beta = np.std(maxima) * np.sqrt(6) / np.pi >>> mu = np.mean(maxima) - 0.57721*beta >>> plt.plot(bins, (1/beta)*np.exp(-(bins - mu)/beta) ... * np.exp(-np.exp(-(bins - mu)/beta)), ... linewidth=2, color='r') >>> plt.plot(bins, 1/(beta * np.sqrt(2 * np.pi)) ... * np.exp(-(bins - mu)**2 / (2 * beta**2)), ... linewidth=2, color='g') >>> plt.show() laplace(loc=0.0, scale=1.0, size=None) Draw samples from the Laplace or double exponential distribution with specified location (or mean) and scale (decay). The Laplace distribution is similar to the Gaussian/normal distribution, but is sharper at the peak and has fatter tails. It represents the difference between two independent, identically distributed exponential random variables. Parameters ---------- loc : float, optional The position, :math:`\mu`, of the distribution peak. scale : float, optional :math:`\lambda`, the exponential decay. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or float Notes ----- It has the probability density function .. math:: f(x; \mu, \lambda) = \frac{1}{2\lambda} \exp\left(-\frac{|x - \mu|}{\lambda}\right). The first law of Laplace, from 1774, states that the frequency of an error can be expressed as an exponential function of the absolute magnitude of the error, which leads to the Laplace distribution. For many problems in economics and health sciences, this distribution seems to model the data better than the standard Gaussian distribution. References ---------- .. [1] Abramowitz, M. and Stegun, I. A. (Eds.). "Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, 9th printing," New York: Dover, 1972. .. [2] Kotz, Samuel, et. al. "The Laplace Distribution and Generalizations, " Birkhauser, 2001. .. [3] Weisstein, Eric W. "Laplace Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/LaplaceDistribution.html .. [4] Wikipedia, "Laplace Distribution", http://en.wikipedia.org/wiki/Laplace_distribution Examples -------- Draw samples from the distribution >>> loc, scale = 0., 1. >>> s = np.random.laplace(loc, scale, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 30, normed=True) >>> x = np.arange(-8., 8., .01) >>> pdf = np.exp(-abs(x-loc)/scale)/(2.*scale) >>> plt.plot(x, pdf) Plot Gaussian for comparison: >>> g = (1/(scale * np.sqrt(2 * np.pi)) * ... np.exp(-(x - loc)**2 / (2 * scale**2))) >>> plt.plot(x,g) power(a, size=None) Draws samples in [0, 1] from a power distribution with positive exponent a - 1. Also known as the power function distribution. Parameters ---------- a : float parameter, > 0 size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar The returned samples lie in [0, 1]. Raises ------ ValueError If a < 1. Notes ----- The probability density function is .. math:: P(x; a) = ax^{a-1}, 0 \le x \le 1, a>0. The power function distribution is just the inverse of the Pareto distribution. It may also be seen as a special case of the Beta distribution. It is used, for example, in modeling the over-reporting of insurance claims. References ---------- .. [1] Christian Kleiber, Samuel Kotz, "Statistical size distributions in economics and actuarial sciences", Wiley, 2003. .. [2] Heckert, N. A. and Filliben, James J. "NIST Handbook 148: Dataplot Reference Manual, Volume 2: Let Subcommands and Library Functions", National Institute of Standards and Technology Handbook Series, June 2003. http://www.itl.nist.gov/div898/software/dataplot/refman2/auxillar/powpdf.pdf Examples -------- Draw samples from the distribution: >>> a = 5. # shape >>> samples = 1000 >>> s = np.random.power(a, samples) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, bins=30) >>> x = np.linspace(0, 1, 100) >>> y = a*x**(a-1.) >>> normed_y = samples*np.diff(bins)[0]*y >>> plt.plot(x, normed_y) >>> plt.show() Compare the power function distribution to the inverse of the Pareto. >>> from scipy import stats >>> rvs = np.random.power(5, 1000000) >>> rvsp = np.random.pareto(5, 1000000) >>> xx = np.linspace(0,1,100) >>> powpdf = stats.powerlaw.pdf(xx,5) >>> plt.figure() >>> plt.hist(rvs, bins=50, normed=True) >>> plt.plot(xx,powpdf,'r-') >>> plt.title('np.random.power(5)') >>> plt.figure() >>> plt.hist(1./(1.+rvsp), bins=50, normed=True) >>> plt.plot(xx,powpdf,'r-') >>> plt.title('inverse of 1 + np.random.pareto(5)') >>> plt.figure() >>> plt.hist(1./(1.+rvsp), bins=50, normed=True) >>> plt.plot(xx,powpdf,'r-') >>> plt.title('inverse of stats.pareto(5)') weibull(a, size=None) Draw samples from a Weibull distribution. Draw samples from a 1-parameter Weibull distribution with the given shape parameter `a`. .. math:: X = (-ln(U))^{1/a} Here, U is drawn from the uniform distribution over (0,1]. The more common 2-parameter Weibull, including a scale parameter :math:`\lambda` is just :math:`X = \lambda(-ln(U))^{1/a}`. Parameters ---------- a : float Shape of the distribution. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray See Also -------- scipy.stats.distributions.weibull_max scipy.stats.distributions.weibull_min scipy.stats.distributions.genextreme gumbel Notes ----- The Weibull (or Type III asymptotic extreme value distribution for smallest values, SEV Type III, or Rosin-Rammler distribution) is one of a class of Generalized Extreme Value (GEV) distributions used in modeling extreme value problems. This class includes the Gumbel and Frechet distributions. The probability density for the Weibull distribution is .. math:: p(x) = \frac{a} {\lambda}(\frac{x}{\lambda})^{a-1}e^{-(x/\lambda)^a}, where :math:`a` is the shape and :math:`\lambda` the scale. The function has its peak (the mode) at :math:`\lambda(\frac{a-1}{a})^{1/a}`. When ``a = 1``, the Weibull distribution reduces to the exponential distribution. References ---------- .. [1] Waloddi Weibull, Royal Technical University, Stockholm, 1939 "A Statistical Theory Of The Strength Of Materials", Ingeniorsvetenskapsakademiens Handlingar Nr 151, 1939, Generalstabens Litografiska Anstalts Forlag, Stockholm. .. [2] Waloddi Weibull, "A Statistical Distribution Function of Wide Applicability", Journal Of Applied Mechanics ASME Paper 1951. .. [3] Wikipedia, "Weibull distribution", http://en.wikipedia.org/wiki/Weibull_distribution Examples -------- Draw samples from the distribution: >>> a = 5. # shape >>> s = np.random.weibull(a, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> x = np.arange(1,100.)/50. >>> def weib(x,n,a): ... return (a / n) * (x / n)**(a - 1) * np.exp(-(x / n)**a) >>> count, bins, ignored = plt.hist(np.random.weibull(5.,1000)) >>> x = np.arange(1,100.)/50. >>> scale = count.max()/weib(x, 1., 5.).max() >>> plt.plot(x, weib(x, 1., 5.)*scale) >>> plt.show() pareto(a, size=None) Draw samples from a Pareto II or Lomax distribution with specified shape. The Lomax or Pareto II distribution is a shifted Pareto distribution. The classical Pareto distribution can be obtained from the Lomax distribution by adding 1 and multiplying by the scale parameter ``m`` (see Notes). The smallest value of the Lomax distribution is zero while for the classical Pareto distribution it is ``mu``, where the standard Pareto distribution has location ``mu = 1``. Lomax can also be considered as a simplified version of the Generalized Pareto distribution (available in SciPy), with the scale set to one and the location set to zero. The Pareto distribution must be greater than zero, and is unbounded above. It is also known as the "80-20 rule". In this distribution, 80 percent of the weights are in the lowest 20 percent of the range, while the other 20 percent fill the remaining 80 percent of the range. Parameters ---------- shape : float, > 0. Shape of the distribution. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. See Also -------- scipy.stats.distributions.lomax.pdf : probability density function, distribution or cumulative density function, etc. scipy.stats.distributions.genpareto.pdf : probability density function, distribution or cumulative density function, etc. Notes ----- The probability density for the Pareto distribution is .. math:: p(x) = \frac{am^a}{x^{a+1}} where :math:`a` is the shape and :math:`m` the scale. The Pareto distribution, named after the Italian economist Vilfredo Pareto, is a power law probability distribution useful in many real world problems. Outside the field of economics it is generally referred to as the Bradford distribution. Pareto developed the distribution to describe the distribution of wealth in an economy. It has also found use in insurance, web page access statistics, oil field sizes, and many other problems, including the download frequency for projects in Sourceforge [1]_. It is one of the so-called "fat-tailed" distributions. References ---------- .. [1] Francis Hunt and Paul Johnson, On the Pareto Distribution of Sourceforge projects. .. [2] Pareto, V. (1896). Course of Political Economy. Lausanne. .. [3] Reiss, R.D., Thomas, M.(2001), Statistical Analysis of Extreme Values, Birkhauser Verlag, Basel, pp 23-30. .. [4] Wikipedia, "Pareto distribution", http://en.wikipedia.org/wiki/Pareto_distribution Examples -------- Draw samples from the distribution: >>> a, m = 3., 2. # shape and mode >>> s = (np.random.pareto(a, 1000) + 1) * m Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, _ = plt.hist(s, 100, normed=True) >>> fit = a*m**a / bins**(a+1) >>> plt.plot(bins, max(count)*fit/max(fit), linewidth=2, color='r') >>> plt.show() vonmises(mu, kappa, size=None) Draw samples from a von Mises distribution. Samples are drawn from a von Mises distribution with specified mode (mu) and dispersion (kappa), on the interval [-pi, pi]. The von Mises distribution (also known as the circular normal distribution) is a continuous probability distribution on the unit circle. It may be thought of as the circular analogue of the normal distribution. Parameters ---------- mu : float Mode ("center") of the distribution. kappa : float Dispersion of the distribution, has to be >=0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : scalar or ndarray The returned samples, which are in the interval [-pi, pi]. See Also -------- scipy.stats.distributions.vonmises : probability density function, distribution, or cumulative density function, etc. Notes ----- The probability density for the von Mises distribution is .. math:: p(x) = \frac{e^{\kappa cos(x-\mu)}}{2\pi I_0(\kappa)}, where :math:`\mu` is the mode and :math:`\kappa` the dispersion, and :math:`I_0(\kappa)` is the modified Bessel function of order 0. The von Mises is named for Richard Edler von Mises, who was born in Austria-Hungary, in what is now the Ukraine. He fled to the United States in 1939 and became a professor at Harvard. He worked in probability theory, aerodynamics, fluid mechanics, and philosophy of science. References ---------- .. [1] Abramowitz, M. and Stegun, I. A. (Eds.). "Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, 9th printing," New York: Dover, 1972. .. [2] von Mises, R., "Mathematical Theory of Probability and Statistics", New York: Academic Press, 1964. Examples -------- Draw samples from the distribution: >>> mu, kappa = 0.0, 4.0 # mean and dispersion >>> s = np.random.vonmises(mu, kappa, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> from scipy.special import i0 >>> plt.hist(s, 50, normed=True) >>> x = np.linspace(-np.pi, np.pi, num=51) >>> y = np.exp(kappa*np.cos(x-mu))/(2*np.pi*i0(kappa)) >>> plt.plot(x, y, linewidth=2, color='r') >>> plt.show() standard_t(df, size=None) Draw samples from a standard Student's t distribution with `df` degrees of freedom. A special case of the hyperbolic distribution. As `df` gets large, the result resembles that of the standard normal distribution (`standard_normal`). Parameters ---------- df : int Degrees of freedom, should be > 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar Drawn samples. Notes ----- The probability density function for the t distribution is .. math:: P(x, df) = \frac{\Gamma(\frac{df+1}{2})}{\sqrt{\pi df} \Gamma(\frac{df}{2})}\Bigl( 1+\frac{x^2}{df} \Bigr)^{-(df+1)/2} The t test is based on an assumption that the data come from a Normal distribution. The t test provides a way to test whether the sample mean (that is the mean calculated from the data) is a good estimate of the true mean. The derivation of the t-distribution was first published in 1908 by William Gisset while working for the Guinness Brewery in Dublin. Due to proprietary issues, he had to publish under a pseudonym, and so he used the name Student. References ---------- .. [1] Dalgaard, Peter, "Introductory Statistics With R", Springer, 2002. .. [2] Wikipedia, "Student's t-distribution" http://en.wikipedia.org/wiki/Student's_t-distribution Examples -------- From Dalgaard page 83 [1]_, suppose the daily energy intake for 11 women in Kj is: >>> intake = np.array([5260., 5470, 5640, 6180, 6390, 6515, 6805, 7515, \ ... 7515, 8230, 8770]) Does their energy intake deviate systematically from the recommended value of 7725 kJ? We have 10 degrees of freedom, so is the sample mean within 95% of the recommended value? >>> s = np.random.standard_t(10, size=100000) >>> np.mean(intake) 6753.636363636364 >>> intake.std(ddof=1) 1142.1232221373727 Calculate the t statistic, setting the ddof parameter to the unbiased value so the divisor in the standard deviation will be degrees of freedom, N-1. >>> t = (np.mean(intake)-7725)/(intake.std(ddof=1)/np.sqrt(len(intake))) >>> import matplotlib.pyplot as plt >>> h = plt.hist(s, bins=100, normed=True) For a one-sided t-test, how far out in the distribution does the t statistic appear? >>> np.sum(s>> s = np.random.standard_cauchy(1000000) >>> s = s[(s>-25) & (s<25)] # truncate distribution so it plots well >>> plt.hist(s, bins=100) >>> plt.show() noncentral_chisquare(df, nonc, size=None) Draw samples from a noncentral chi-square distribution. The noncentral :math:`\chi^2` distribution is a generalisation of the :math:`\chi^2` distribution. Parameters ---------- df : int Degrees of freedom, should be > 0 as of Numpy 1.10, should be > 1 for earlier versions. nonc : float Non-centrality, should be non-negative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Notes ----- The probability density function for the noncentral Chi-square distribution is .. math:: P(x;df,nonc) = \sum^{\infty}_{i=0} \frac{e^{-nonc/2}(nonc/2)^{i}}{i!} \P_{Y_{df+2i}}(x), where :math:`Y_{q}` is the Chi-square with q degrees of freedom. In Delhi (2007), it is noted that the noncentral chi-square is useful in bombing and coverage problems, the probability of killing the point target given by the noncentral chi-squared distribution. References ---------- .. [1] Delhi, M.S. Holla, "On a noncentral chi-square distribution in the analysis of weapon systems effectiveness", Metrika, Volume 15, Number 1 / December, 1970. .. [2] Wikipedia, "Noncentral chi-square distribution" http://en.wikipedia.org/wiki/Noncentral_chi-square_distribution Examples -------- Draw values from the distribution and plot the histogram >>> import matplotlib.pyplot as plt >>> values = plt.hist(np.random.noncentral_chisquare(3, 20, 100000), ... bins=200, normed=True) >>> plt.show() Draw values from a noncentral chisquare with very small noncentrality, and compare to a chisquare. >>> plt.figure() >>> values = plt.hist(np.random.noncentral_chisquare(3, .0000001, 100000), ... bins=np.arange(0., 25, .1), normed=True) >>> values2 = plt.hist(np.random.chisquare(3, 100000), ... bins=np.arange(0., 25, .1), normed=True) >>> plt.plot(values[1][0:-1], values[0]-values2[0], 'ob') >>> plt.show() Demonstrate how large values of non-centrality lead to a more symmetric distribution. >>> plt.figure() >>> values = plt.hist(np.random.noncentral_chisquare(3, 20, 100000), ... bins=200, normed=True) >>> plt.show() chisquare(df, size=None) Draw samples from a chi-square distribution. When `df` independent random variables, each with standard normal distributions (mean 0, variance 1), are squared and summed, the resulting distribution is chi-square (see Notes). This distribution is often used in hypothesis testing. Parameters ---------- df : int Number of degrees of freedom. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- output : ndarray Samples drawn from the distribution, packed in a `size`-shaped array. Raises ------ ValueError When `df` <= 0 or when an inappropriate `size` (e.g. ``size=-1``) is given. Notes ----- The variable obtained by summing the squares of `df` independent, standard normally distributed random variables: .. math:: Q = \sum_{i=0}^{\mathtt{df}} X^2_i is chi-square distributed, denoted .. math:: Q \sim \chi^2_k. The probability density function of the chi-squared distribution is .. math:: p(x) = \frac{(1/2)^{k/2}}{\Gamma(k/2)} x^{k/2 - 1} e^{-x/2}, where :math:`\Gamma` is the gamma function, .. math:: \Gamma(x) = \int_0^{-\infty} t^{x - 1} e^{-t} dt. References ---------- .. [1] NIST "Engineering Statistics Handbook" http://www.itl.nist.gov/div898/handbook/eda/section3/eda3666.htm Examples -------- >>> np.random.chisquare(2,4) array([ 1.89920014, 9.00867716, 3.13710533, 5.62318272]) noncentral_f(dfnum, dfden, nonc, size=None) Draw samples from the noncentral F distribution. Samples are drawn from an F distribution with specified parameters, `dfnum` (degrees of freedom in numerator) and `dfden` (degrees of freedom in denominator), where both parameters > 1. `nonc` is the non-centrality parameter. Parameters ---------- dfnum : int Parameter, should be > 1. dfden : int Parameter, should be > 1. nonc : float Parameter, should be >= 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : scalar or ndarray Drawn samples. Notes ----- When calculating the power of an experiment (power = probability of rejecting the null hypothesis when a specific alternative is true) the non-central F statistic becomes important. When the null hypothesis is true, the F statistic follows a central F distribution. When the null hypothesis is not true, then it follows a non-central F statistic. References ---------- .. [1] Weisstein, Eric W. "Noncentral F-Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/NoncentralF-Distribution.html .. [2] Wikipedia, "Noncentral F distribution", http://en.wikipedia.org/wiki/Noncentral_F-distribution Examples -------- In a study, testing for a specific alternative to the null hypothesis requires use of the Noncentral F distribution. We need to calculate the area in the tail of the distribution that exceeds the value of the F distribution for the null hypothesis. We'll plot the two probability distributions for comparison. >>> dfnum = 3 # between group deg of freedom >>> dfden = 20 # within groups degrees of freedom >>> nonc = 3.0 >>> nc_vals = np.random.noncentral_f(dfnum, dfden, nonc, 1000000) >>> NF = np.histogram(nc_vals, bins=50, normed=True) >>> c_vals = np.random.f(dfnum, dfden, 1000000) >>> F = np.histogram(c_vals, bins=50, normed=True) >>> plt.plot(F[1][1:], F[0]) >>> plt.plot(NF[1][1:], NF[0]) >>> plt.show() f(dfnum, dfden, size=None) Draw samples from an F distribution. Samples are drawn from an F distribution with specified parameters, `dfnum` (degrees of freedom in numerator) and `dfden` (degrees of freedom in denominator), where both parameters should be greater than zero. The random variate of the F distribution (also known as the Fisher distribution) is a continuous probability distribution that arises in ANOVA tests, and is the ratio of two chi-square variates. Parameters ---------- dfnum : float Degrees of freedom in numerator. Should be greater than zero. dfden : float Degrees of freedom in denominator. Should be greater than zero. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar Samples from the Fisher distribution. See Also -------- scipy.stats.distributions.f : probability density function, distribution or cumulative density function, etc. Notes ----- The F statistic is used to compare in-group variances to between-group variances. Calculating the distribution depends on the sampling, and so it is a function of the respective degrees of freedom in the problem. The variable `dfnum` is the number of samples minus one, the between-groups degrees of freedom, while `dfden` is the within-groups degrees of freedom, the sum of the number of samples in each group minus the number of groups. References ---------- .. [1] Glantz, Stanton A. "Primer of Biostatistics.", McGraw-Hill, Fifth Edition, 2002. .. [2] Wikipedia, "F-distribution", http://en.wikipedia.org/wiki/F-distribution Examples -------- An example from Glantz[1], pp 47-40: Two groups, children of diabetics (25 people) and children from people without diabetes (25 controls). Fasting blood glucose was measured, case group had a mean value of 86.1, controls had a mean value of 82.2. Standard deviations were 2.09 and 2.49 respectively. Are these data consistent with the null hypothesis that the parents diabetic status does not affect their children's blood glucose levels? Calculating the F statistic from the data gives a value of 36.01. Draw samples from the distribution: >>> dfnum = 1. # between group degrees of freedom >>> dfden = 48. # within groups degrees of freedom >>> s = np.random.f(dfnum, dfden, 1000) The lower bound for the top 1% of the samples is : >>> sort(s)[-10] 7.61988120985 So there is about a 1% chance that the F statistic will exceed 7.62, the measured value is 36, so the null hypothesis is rejected at the 1% level. gamma(shape, scale=1.0, size=None) Draw samples from a Gamma distribution. Samples are drawn from a Gamma distribution with specified parameters, `shape` (sometimes designated "k") and `scale` (sometimes designated "theta"), where both parameters are > 0. Parameters ---------- shape : scalar > 0 The shape of the gamma distribution. scale : scalar > 0, optional The scale of the gamma distribution. Default is equal to 1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : ndarray, float Returns one sample unless `size` parameter is specified. See Also -------- scipy.stats.distributions.gamma : probability density function, distribution or cumulative density function, etc. Notes ----- The probability density for the Gamma distribution is .. math:: p(x) = x^{k-1}\frac{e^{-x/\theta}}{\theta^k\Gamma(k)}, where :math:`k` is the shape and :math:`\theta` the scale, and :math:`\Gamma` is the Gamma function. The Gamma distribution is often used to model the times to failure of electronic components, and arises naturally in processes for which the waiting times between Poisson distributed events are relevant. References ---------- .. [1] Weisstein, Eric W. "Gamma Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/GammaDistribution.html .. [2] Wikipedia, "Gamma-distribution", http://en.wikipedia.org/wiki/Gamma-distribution Examples -------- Draw samples from the distribution: >>> shape, scale = 2., 2. # mean and dispersion >>> s = np.random.gamma(shape, scale, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> import scipy.special as sps >>> count, bins, ignored = plt.hist(s, 50, normed=True) >>> y = bins**(shape-1)*(np.exp(-bins/scale) / ... (sps.gamma(shape)*scale**shape)) >>> plt.plot(bins, y, linewidth=2, color='r') >>> plt.show() standard_gamma(shape, size=None) Draw samples from a standard Gamma distribution. Samples are drawn from a Gamma distribution with specified parameters, shape (sometimes designated "k") and scale=1. Parameters ---------- shape : float Parameter, should be > 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar The drawn samples. See Also -------- scipy.stats.distributions.gamma : probability density function, distribution or cumulative density function, etc. Notes ----- The probability density for the Gamma distribution is .. math:: p(x) = x^{k-1}\frac{e^{-x/\theta}}{\theta^k\Gamma(k)}, where :math:`k` is the shape and :math:`\theta` the scale, and :math:`\Gamma` is the Gamma function. The Gamma distribution is often used to model the times to failure of electronic components, and arises naturally in processes for which the waiting times between Poisson distributed events are relevant. References ---------- .. [1] Weisstein, Eric W. "Gamma Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/GammaDistribution.html .. [2] Wikipedia, "Gamma-distribution", http://en.wikipedia.org/wiki/Gamma-distribution Examples -------- Draw samples from the distribution: >>> shape, scale = 2., 1. # mean and width >>> s = np.random.standard_gamma(shape, 1000000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> import scipy.special as sps >>> count, bins, ignored = plt.hist(s, 50, normed=True) >>> y = bins**(shape-1) * ((np.exp(-bins/scale))/ \ ... (sps.gamma(shape) * scale**shape)) >>> plt.plot(bins, y, linewidth=2, color='r') >>> plt.show() standard_exponential(size=None) Draw samples from the standard exponential distribution. `standard_exponential` is identical to the exponential distribution with a scale parameter of 1. Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : float or ndarray Drawn samples. Examples -------- Output a 3x8000 array: >>> n = np.random.standard_exponential((3, 8000)) exponential(scale=1.0, size=None) Draw samples from an exponential distribution. Its probability density function is .. math:: f(x; \frac{1}{\beta}) = \frac{1}{\beta} \exp(-\frac{x}{\beta}), for ``x > 0`` and 0 elsewhere. :math:`\beta` is the scale parameter, which is the inverse of the rate parameter :math:`\lambda = 1/\beta`. The rate parameter is an alternative, widely used parameterization of the exponential distribution [3]_. The exponential distribution is a continuous analogue of the geometric distribution. It describes many common situations, such as the size of raindrops measured over many rainstorms [1]_, or the time between page requests to Wikipedia [2]_. Parameters ---------- scale : float The scale parameter, :math:`\beta = 1/\lambda`. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. References ---------- .. [1] Peyton Z. Peebles Jr., "Probability, Random Variables and Random Signal Principles", 4th ed, 2001, p. 57. .. [2] "Poisson Process", Wikipedia, http://en.wikipedia.org/wiki/Poisson_process .. [3] "Exponential Distribution, Wikipedia, http://en.wikipedia.org/wiki/Exponential_distribution beta(a, b, size=None) Draw samples from a Beta distribution. The Beta distribution is a special case of the Dirichlet distribution, and is related to the Gamma distribution. It has the probability distribution function .. math:: f(x; a,b) = \frac{1}{B(\alpha, \beta)} x^{\alpha - 1} (1 - x)^{\beta - 1}, where the normalisation, B, is the beta function, .. math:: B(\alpha, \beta) = \int_0^1 t^{\alpha - 1} (1 - t)^{\beta - 1} dt. It is often seen in Bayesian inference and order statistics. Parameters ---------- a : float Alpha, non-negative. b : float Beta, non-negative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : ndarray Array of the given shape, containing values drawn from a Beta distribution. normal(loc=0.0, scale=1.0, size=None) Draw random samples from a normal (Gaussian) distribution. The probability density function of the normal distribution, first derived by De Moivre and 200 years later by both Gauss and Laplace independently [2]_, is often called the bell curve because of its characteristic shape (see the example below). The normal distributions occurs often in nature. For example, it describes the commonly occurring distribution of samples influenced by a large number of tiny, random disturbances, each with its own unique distribution [2]_. Parameters ---------- loc : float Mean ("centre") of the distribution. scale : float Standard deviation (spread or "width") of the distribution. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. See Also -------- scipy.stats.distributions.norm : probability density function, distribution or cumulative density function, etc. Notes ----- The probability density for the Gaussian distribution is .. math:: p(x) = \frac{1}{\sqrt{ 2 \pi \sigma^2 }} e^{ - \frac{ (x - \mu)^2 } {2 \sigma^2} }, where :math:`\mu` is the mean and :math:`\sigma` the standard deviation. The square of the standard deviation, :math:`\sigma^2`, is called the variance. The function has its peak at the mean, and its "spread" increases with the standard deviation (the function reaches 0.607 times its maximum at :math:`x + \sigma` and :math:`x - \sigma` [2]_). This implies that `numpy.random.normal` is more likely to return samples lying close to the mean, rather than those far away. References ---------- .. [1] Wikipedia, "Normal distribution", http://en.wikipedia.org/wiki/Normal_distribution .. [2] P. R. Peebles Jr., "Central Limit Theorem" in "Probability, Random Variables and Random Signal Principles", 4th ed., 2001, pp. 51, 51, 125. Examples -------- Draw samples from the distribution: >>> mu, sigma = 0, 0.1 # mean and standard deviation >>> s = np.random.normal(mu, sigma, 1000) Verify the mean and the variance: >>> abs(mu - np.mean(s)) < 0.01 True >>> abs(sigma - np.std(s, ddof=1)) < 0.01 True Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 30, normed=True) >>> plt.plot(bins, 1/(sigma * np.sqrt(2 * np.pi)) * ... np.exp( - (bins - mu)**2 / (2 * sigma**2) ), ... linewidth=2, color='r') >>> plt.show() standard_normal(size=None) Draw samples from a standard Normal distribution (mean=0, stdev=1). Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : float or ndarray Drawn samples. Examples -------- >>> s = np.random.standard_normal(8000) >>> s array([ 0.6888893 , 0.78096262, -0.89086505, ..., 0.49876311, #random -0.38672696, -0.4685006 ]) #random >>> s.shape (8000,) >>> s = np.random.standard_normal(size=(3, 4, 2)) >>> s.shape (3, 4, 2) random_integers(low, high=None, size=None) Random integers of type np.int between `low` and `high`, inclusive. Return random integers of type np.int from the "discrete uniform" distribution in the closed interval [`low`, `high`]. If `high` is None (the default), then results are from [1, `low`]. The np.int type translates to the C long type used by Python 2 for "short" integers and its precision is platform dependent. This function has been deprecated. Use randint instead. .. deprecated:: 1.11.0 Parameters ---------- low : int Lowest (signed) integer to be drawn from the distribution (unless ``high=None``, in which case this parameter is the *highest* such integer). high : int, optional If provided, the largest (signed) integer to be drawn from the distribution (see above for behavior if ``high=None``). size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : int or ndarray of ints `size`-shaped array of random integers from the appropriate distribution, or a single such random int if `size` not provided. See Also -------- random.randint : Similar to `random_integers`, only for the half-open interval [`low`, `high`), and 0 is the lowest value if `high` is omitted. Notes ----- To sample from N evenly spaced floating-point numbers between a and b, use:: a + (b - a) * (np.random.random_integers(N) - 1) / (N - 1.) Examples -------- >>> np.random.random_integers(5) 4 >>> type(np.random.random_integers(5)) >>> np.random.random_integers(5, size=(3.,2.)) array([[5, 4], [3, 3], [4, 5]]) Choose five random numbers from the set of five evenly-spaced numbers between 0 and 2.5, inclusive (*i.e.*, from the set :math:`{0, 5/8, 10/8, 15/8, 20/8}`): >>> 2.5 * (np.random.random_integers(5, size=(5,)) - 1) / 4. array([ 0.625, 1.25 , 0.625, 0.625, 2.5 ]) Roll two six sided dice 1000 times and sum the results: >>> d1 = np.random.random_integers(1, 6, 1000) >>> d2 = np.random.random_integers(1, 6, 1000) >>> dsums = d1 + d2 Display results as a histogram: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(dsums, 11, normed=True) >>> plt.show() randn(d0, d1, ..., dn) Return a sample (or samples) from the "standard normal" distribution. If positive, int_like or int-convertible arguments are provided, `randn` generates an array of shape ``(d0, d1, ..., dn)``, filled with random floats sampled from a univariate "normal" (Gaussian) distribution of mean 0 and variance 1 (if any of the :math:`d_i` are floats, they are first converted to integers by truncation). A single float randomly sampled from the distribution is returned if no argument is provided. This is a convenience function. If you want an interface that takes a tuple as the first argument, use `numpy.random.standard_normal` instead. Parameters ---------- d0, d1, ..., dn : int, optional The dimensions of the returned array, should be all positive. If no argument is given a single Python float is returned. Returns ------- Z : ndarray or float A ``(d0, d1, ..., dn)``-shaped array of floating-point samples from the standard normal distribution, or a single such float if no parameters were supplied. See Also -------- random.standard_normal : Similar, but takes a tuple as its argument. Notes ----- For random samples from :math:`N(\mu, \sigma^2)`, use: ``sigma * np.random.randn(...) + mu`` Examples -------- >>> np.random.randn() 2.1923875335537315 #random Two-by-four array of samples from N(3, 6.25): >>> 2.5 * np.random.randn(2, 4) + 3 array([[-4.49401501, 4.00950034, -1.81814867, 7.29718677], #random [ 0.39924804, 4.68456316, 4.99394529, 4.84057254]]) #random rand(d0, d1, ..., dn) Random values in a given shape. Create an array of the given shape and populate it with random samples from a uniform distribution over ``[0, 1)``. Parameters ---------- d0, d1, ..., dn : int, optional The dimensions of the returned array, should all be positive. If no argument is given a single Python float is returned. Returns ------- out : ndarray, shape ``(d0, d1, ..., dn)`` Random values. See Also -------- random Notes ----- This is a convenience function. If you want an interface that takes a shape-tuple as the first argument, refer to np.random.random_sample . Examples -------- >>> np.random.rand(3,2) array([[ 0.14022471, 0.96360618], #random [ 0.37601032, 0.25528411], #random [ 0.49313049, 0.94909878]]) #random uniform(low=0.0, high=1.0, size=None) Draw samples from a uniform distribution. Samples are uniformly distributed over the half-open interval ``[low, high)`` (includes low, but excludes high). In other words, any value within the given interval is equally likely to be drawn by `uniform`. Parameters ---------- low : float, optional Lower boundary of the output interval. All values generated will be greater than or equal to low. The default value is 0. high : float Upper boundary of the output interval. All values generated will be less than high. The default value is 1.0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : ndarray Drawn samples, with shape `size`. See Also -------- randint : Discrete uniform distribution, yielding integers. random_integers : Discrete uniform distribution over the closed interval ``[low, high]``. random_sample : Floats uniformly distributed over ``[0, 1)``. random : Alias for `random_sample`. rand : Convenience function that accepts dimensions as input, e.g., ``rand(2,2)`` would generate a 2-by-2 array of floats, uniformly distributed over ``[0, 1)``. Notes ----- The probability density function of the uniform distribution is .. math:: p(x) = \frac{1}{b - a} anywhere within the interval ``[a, b)``, and zero elsewhere. Examples -------- Draw samples from the distribution: >>> s = np.random.uniform(-1,0,1000) All values are within the given interval: >>> np.all(s >= -1) True >>> np.all(s < 0) True Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 15, normed=True) >>> plt.plot(bins, np.ones_like(bins), linewidth=2, color='r') >>> plt.show() choice(a, size=None, replace=True, p=None) Generates a random sample from a given 1-D array .. versionadded:: 1.7.0 Parameters ----------- a : 1-D array-like or int If an ndarray, a random sample is generated from its elements. If an int, the random sample is generated as if a was np.arange(n) size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. replace : boolean, optional Whether the sample is with or without replacement p : 1-D array-like, optional The probabilities associated with each entry in a. If not given the sample assumes a uniform distribution over all entries in a. Returns -------- samples : 1-D ndarray, shape (size,) The generated random samples Raises ------- ValueError If a is an int and less than zero, if a or p are not 1-dimensional, if a is an array-like of size 0, if p is not a vector of probabilities, if a and p have different lengths, or if replace=False and the sample size is greater than the population size See Also --------- randint, shuffle, permutation Examples --------- Generate a uniform random sample from np.arange(5) of size 3: >>> np.random.choice(5, 3) array([0, 3, 4]) >>> #This is equivalent to np.random.randint(0,5,3) Generate a non-uniform random sample from np.arange(5) of size 3: >>> np.random.choice(5, 3, p=[0.1, 0, 0.3, 0.6, 0]) array([3, 3, 0]) Generate a uniform random sample from np.arange(5) of size 3 without replacement: >>> np.random.choice(5, 3, replace=False) array([3,1,0]) >>> #This is equivalent to np.random.permutation(np.arange(5))[:3] Generate a non-uniform random sample from np.arange(5) of size 3 without replacement: >>> np.random.choice(5, 3, replace=False, p=[0.1, 0, 0.3, 0.6, 0]) array([2, 3, 0]) Any of the above can be repeated with an arbitrary array-like instead of just integers. For instance: >>> aa_milne_arr = ['pooh', 'rabbit', 'piglet', 'Christopher'] >>> np.random.choice(aa_milne_arr, 5, p=[0.5, 0.1, 0.1, 0.3]) array(['pooh', 'pooh', 'pooh', 'Christopher', 'piglet'], dtype='|S11') bytes(length) Return random bytes. Parameters ---------- length : int Number of random bytes. Returns ------- out : str String of length `length`. Examples -------- >>> np.random.bytes(10) ' eh\x85\x022SZ\xbf\xa4' #random randint(low, high=None, size=None, dtype='l') Return random integers from `low` (inclusive) to `high` (exclusive). Return random integers from the "discrete uniform" distribution of the specified dtype in the "half-open" interval [`low`, `high`). If `high` is None (the default), then results are from [0, `low`). Parameters ---------- low : int Lowest (signed) integer to be drawn from the distribution (unless ``high=None``, in which case this parameter is the *highest* such integer). high : int, optional If provided, one above the largest (signed) integer to be drawn from the distribution (see above for behavior if ``high=None``). size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. dtype : dtype, optional Desired dtype of the result. All dtypes are determined by their name, i.e., 'int64', 'int', etc, so byteorder is not available and a specific precision may have different C types depending on the platform. The default value is 'np.int'. .. versionadded:: 1.11.0 Returns ------- out : int or ndarray of ints `size`-shaped array of random integers from the appropriate distribution, or a single such random int if `size` not provided. See Also -------- random.random_integers : similar to `randint`, only for the closed interval [`low`, `high`], and 1 is the lowest value if `high` is omitted. In particular, this other one is the one to use to generate uniformly distributed discrete non-integers. Examples -------- >>> np.random.randint(2, size=10) array([1, 0, 0, 0, 1, 1, 0, 0, 1, 0]) >>> np.random.randint(1, size=10) array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0]) Generate a 2 x 4 array of ints between 0 and 4, inclusive: >>> np.random.randint(5, size=(2, 4)) array([[4, 0, 2, 1], [3, 2, 2, 0]]) tomaxint(size=None) Random integers between 0 and ``sys.maxint``, inclusive. Return a sample of uniformly distributed random integers in the interval [0, ``sys.maxint``]. Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : ndarray Drawn samples, with shape `size`. See Also -------- randint : Uniform sampling over a given half-open interval of integers. random_integers : Uniform sampling over a given closed interval of integers. Examples -------- >>> RS = np.random.mtrand.RandomState() # need a RandomState object >>> RS.tomaxint((2,2,2)) array([[[1170048599, 1600360186], [ 739731006, 1947757578]], [[1871712945, 752307660], [1601631370, 1479324245]]]) >>> import sys >>> sys.maxint 2147483647 >>> RS.tomaxint((2,2,2)) < sys.maxint array([[[ True, True], [ True, True]], [[ True, True], [ True, True]]], dtype=bool) random_sample(size=None) Return random floats in the half-open interval [0.0, 1.0). Results are from the "continuous uniform" distribution over the stated interval. To sample :math:`Unif[a, b), b > a` multiply the output of `random_sample` by `(b-a)` and add `a`:: (b - a) * random_sample() + a Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : float or ndarray of floats Array of random floats of shape `size` (unless ``size=None``, in which case a single float is returned). Examples -------- >>> np.random.random_sample() 0.47108547995356098 >>> type(np.random.random_sample()) >>> np.random.random_sample((5,)) array([ 0.30220482, 0.86820401, 0.1654503 , 0.11659149, 0.54323428]) Three-by-two array of random numbers from [-5, 0): >>> 5 * np.random.random_sample((3, 2)) - 5 array([[-3.99149989, -0.52338984], [-2.99091858, -0.79479508], [-1.23204345, -1.75224494]]) set_state(state) Set the internal state of the generator from a tuple. For use if one has reason to manually (re-)set the internal state of the "Mersenne Twister"[1]_ pseudo-random number generating algorithm. Parameters ---------- state : tuple(str, ndarray of 624 uints, int, int, float) The `state` tuple has the following items: 1. the string 'MT19937', specifying the Mersenne Twister algorithm. 2. a 1-D array of 624 unsigned integers ``keys``. 3. an integer ``pos``. 4. an integer ``has_gauss``. 5. a float ``cached_gaussian``. Returns ------- out : None Returns 'None' on success. See Also -------- get_state Notes ----- `set_state` and `get_state` are not needed to work with any of the random distributions in NumPy. If the internal state is manually altered, the user should know exactly what he/she is doing. For backwards compatibility, the form (str, array of 624 uints, int) is also accepted although it is missing some information about the cached Gaussian value: ``state = ('MT19937', keys, pos)``. References ---------- .. [1] M. Matsumoto and T. Nishimura, "Mersenne Twister: A 623-dimensionally equidistributed uniform pseudorandom number generator," *ACM Trans. on Modeling and Computer Simulation*, Vol. 8, No. 1, pp. 3-30, Jan. 1998. get_state() Return a tuple representing the internal state of the generator. For more details, see `set_state`. Returns ------- out : tuple(str, ndarray of 624 uints, int, int, float) The returned tuple has the following items: 1. the string 'MT19937'. 2. a 1-D array of 624 unsigned integer keys. 3. an integer ``pos``. 4. an integer ``has_gauss``. 5. a float ``cached_gaussian``. See Also -------- set_state Notes ----- `set_state` and `get_state` are not needed to work with any of the random distributions in NumPy. If the internal state is manually altered, the user should know exactly what he/she is doing. seed(seed=None) Seed the generator. This method is called when `RandomState` is initialized. It can be called again to re-seed the generator. For details, see `RandomState`. Parameters ---------- seed : int or array_like, optional Seed for `RandomState`. Must be convertible to 32 bit unsigned integers. See Also -------- RandomState _rand_uint64(low, high, size, rngstate) See `_rand_int32` for documentation, only the return type changes. _rand_uint32(self, low, high, size, rngstate) See `_rand_int32` for documentation, only the return type changes. _rand_uint16(low, high, size, rngstate) See `_rand_int32` for documentation, only the return type changes. _rand_uint8(low, high, size, rngstate) See `_rand_int32` for documentation, only the return type changes. _rand_int64(low, high, size, rngstate) See `_rand_int32` for documentation, only the return type changes. _rand_int32(self, low, high, size, rngstate) Return random np.int32 integers between `low` and `high`, inclusive. Return random integers from the "discrete uniform" distribution in the closed interval [`low`, `high`]. On entry the arguments are presumed to have been validated for size and order for the np.int32 type. Parameters ---------- low : int Lowest (signed) integer to be drawn from the distribution. high : int Highest (signed) integer to be drawn from the distribution. size : int or tuple of ints Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. rngstate : encapsulated pointer to rk_state The specific type depends on the python version. In Python 2 it is a PyCObject, in Python 3 a PyCapsule object. Returns ------- out : python scalar or ndarray of np.int32 `size`-shaped array of random integers from the appropriate distribution, or a single such random int if `size` not provided. _rand_int16(low, high, size, rngstate) See `_rand_int32` for documentation, only the return type changes. _rand_int8(low, high, size, rngstate) See `_rand_int32` for documentation, only the return type changes. _rand_bool(low, high, size, rngstate) See `_rand_int32` for documentation, only the return type changes. multivariate_normal__RandomState_ctorDeprecationWarningnegative_binomial_shape_from_sizestandard_normalstandard_cauchyrandom_integerspoisson_lam_maxdummy_threadingstandard_gammahypergeometricRuntimeWarningrandom_samplegreater_equalcount_nonzero_randint_typeOverflowErrorsearchsortedreturn_indexnoncentral_f_rand_uint64_rand_uint32_rand_uint16permutationmultinomialexponential_rand_uint8_rand_int64_rand_int32_rand_int16ImportErrortriangularstandard_tlogical_orless_equalissubdtypeempty_like_rand_int8_rand_bool__import__ValueErrorthreadingset_statelogserieslognormalget_stategeometricdirichletchisquare__enter__TypeErrorwarningsvonmisessubtractrngstatereversedrayleighoperatorlogisticitemsizefloatingbinomial__test____main____exit__weibulluniformstridesshufflereshapereplacerandintpoissonnsamplendarraylaplaceintegergreaterfloat64castingasarrayMT19937unsafeuniqueuint64uint32uint16reducerandomparetonormalmtrandgumbelformatcumsumctypeschoiceastypearangezerosuint8statesigmashapescalerightravelrangerandnpvalspowernumpyngoodkappaisnanint64int32int16indexiinfogammafinfoequalemptydtypedfnumdfdenbytesbool_arrayalpha_randzipfwarnwalduinttakesqrtsortsizesideseedsaferandprodnoncndimnbadnamemodemeanlonglessleftitemintpint8highdatacopyboolbetaaxisLocksvdsumrngoutoffmaxlowloclamintepsdotcovcntbufanyaddnpmudf permutation(x) Randomly permute a sequence, or return a permuted range. If `x` is a multi-dimensional array, it is only shuffled along its first index. Parameters ---------- x : int or array_like If `x` is an integer, randomly permute ``np.arange(x)``. If `x` is an array, make a copy and shuffle the elements randomly. Returns ------- out : ndarray Permuted sequence or array range. Examples -------- >>> np.random.permutation(10) array([1, 7, 4, 3, 0, 9, 2, 5, 8, 6]) >>> np.random.permutation([1, 4, 9, 12, 15]) array([15, 1, 9, 4, 12]) >>> arr = np.arange(9).reshape((3, 3)) >>> np.random.permutation(arr) array([[6, 7, 8], [0, 1, 2], [3, 4, 5]]) RandomState.permutation (line 5041) shuffle(x) Modify a sequence in-place by shuffling its contents. Parameters ---------- x : array_like The array or list to be shuffled. Returns ------- None Examples -------- >>> arr = np.arange(10) >>> np.random.shuffle(arr) >>> arr [1 7 5 2 9 4 3 6 0 8] This function only shuffles the array along the first index of a multi-dimensional array: >>> arr = np.arange(9).reshape((3, 3)) >>> np.random.shuffle(arr) >>> arr array([[3, 4, 5], [6, 7, 8], [0, 1, 2]]) RandomState.shuffle (line 4957) dirichlet(alpha, size=None) Draw samples from the Dirichlet distribution. Draw `size` samples of dimension k from a Dirichlet distribution. A Dirichlet-distributed random variable can be seen as a multivariate generalization of a Beta distribution. Dirichlet pdf is the conjugate prior of a multinomial in Bayesian inference. Parameters ---------- alpha : array Parameter of the distribution (k dimension for sample of dimension k). size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray, The drawn samples, of shape (size, alpha.ndim). Notes ----- .. math:: X \approx \prod_{i=1}^{k}{x^{\alpha_i-1}_i} Uses the following property for computation: for each dimension, draw a random sample y_i from a standard gamma generator of shape `alpha_i`, then :math:`X = \frac{1}{\sum_{i=1}^k{y_i}} (y_1, \ldots, y_n)` is Dirichlet distributed. References ---------- .. [1] David McKay, "Information Theory, Inference and Learning Algorithms," chapter 23, http://www.inference.phy.cam.ac.uk/mackay/ .. [2] Wikipedia, "Dirichlet distribution", http://en.wikipedia.org/wiki/Dirichlet_distribution Examples -------- Taking an example cited in Wikipedia, this distribution can be used if one wanted to cut strings (each of initial length 1.0) into K pieces with different lengths, where each piece had, on average, a designated average length, but allowing some variation in the relative sizes of the pieces. >>> s = np.random.dirichlet((10, 5, 3), 20).transpose() >>> plt.barh(range(20), s[0]) >>> plt.barh(range(20), s[1], left=s[0], color='g') >>> plt.barh(range(20), s[2], left=s[0]+s[1], color='r') >>> plt.title("Lengths of Strings") RandomState.dirichlet (line 4841) multinomial(n, pvals, size=None) Draw samples from a multinomial distribution. The multinomial distribution is a multivariate generalisation of the binomial distribution. Take an experiment with one of ``p`` possible outcomes. An example of such an experiment is throwing a dice, where the outcome can be 1 through 6. Each sample drawn from the distribution represents `n` such experiments. Its values, ``X_i = [X_0, X_1, ..., X_p]``, represent the number of times the outcome was ``i``. Parameters ---------- n : int Number of experiments. pvals : sequence of floats, length p Probabilities of each of the ``p`` different outcomes. These should sum to 1 (however, the last element is always assumed to account for the remaining probability, as long as ``sum(pvals[:-1]) <= 1)``. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : ndarray The drawn samples, of shape *size*, if that was provided. If not, the shape is ``(N,)``. In other words, each entry ``out[i,j,...,:]`` is an N-dimensional value drawn from the distribution. Examples -------- Throw a dice 20 times: >>> np.random.multinomial(20, [1/6.]*6, size=1) array([[4, 1, 7, 5, 2, 1]]) It landed 4 times on 1, once on 2, etc. Now, throw the dice 20 times, and 20 times again: >>> np.random.multinomial(20, [1/6.]*6, size=2) array([[3, 4, 3, 3, 4, 3], [2, 4, 3, 4, 0, 7]]) For the first run, we threw 3 times 1, 4 times 2, etc. For the second, we threw 2 times 1, 4 times 2, etc. A loaded die is more likely to land on number 6: >>> np.random.multinomial(100, [1/7.]*5 + [2/7.]) array([11, 16, 14, 17, 16, 26]) The probability inputs should be normalized. As an implementation detail, the value of the last entry is ignored and assumed to take up any leftover probability mass, but this should not be relied on. A biased coin which has twice as much weight on one side as on the other should be sampled like so: >>> np.random.multinomial(100, [1.0 / 3, 2.0 / 3]) # RIGHT array([38, 62]) not like: >>> np.random.multinomial(100, [1.0, 2.0]) # WRONG array([100, 0]) RandomState.multinomial (line 4728) multivariate_normal(mean, cov[, size]) Draw random samples from a multivariate normal distribution. The multivariate normal, multinormal or Gaussian distribution is a generalization of the one-dimensional normal distribution to higher dimensions. Such a distribution is specified by its mean and covariance matrix. These parameters are analogous to the mean (average or "center") and variance (standard deviation, or "width," squared) of the one-dimensional normal distribution. Parameters ---------- mean : 1-D array_like, of length N Mean of the N-dimensional distribution. cov : 2-D array_like, of shape (N, N) Covariance matrix of the distribution. It must be symmetric and positive-semidefinite for proper sampling. size : int or tuple of ints, optional Given a shape of, for example, ``(m,n,k)``, ``m*n*k`` samples are generated, and packed in an `m`-by-`n`-by-`k` arrangement. Because each sample is `N`-dimensional, the output shape is ``(m,n,k,N)``. If no shape is specified, a single (`N`-D) sample is returned. Returns ------- out : ndarray The drawn samples, of shape *size*, if that was provided. If not, the shape is ``(N,)``. In other words, each entry ``out[i,j,...,:]`` is an N-dimensional value drawn from the distribution. Notes ----- The mean is a coordinate in N-dimensional space, which represents the location where samples are most likely to be generated. This is analogous to the peak of the bell curve for the one-dimensional or univariate normal distribution. Covariance indicates the level to which two variables vary together. From the multivariate normal distribution, we draw N-dimensional samples, :math:`X = [x_1, x_2, ... x_N]`. The covariance matrix element :math:`C_{ij}` is the covariance of :math:`x_i` and :math:`x_j`. The element :math:`C_{ii}` is the variance of :math:`x_i` (i.e. its "spread"). Instead of specifying the full covariance matrix, popular approximations include: - Spherical covariance (*cov* is a multiple of the identity matrix) - Diagonal covariance (*cov* has non-negative elements, and only on the diagonal) This geometrical property can be seen in two dimensions by plotting generated data-points: >>> mean = [0, 0] >>> cov = [[1, 0], [0, 100]] # diagonal covariance Diagonal covariance means that points are oriented along x or y-axis: >>> import matplotlib.pyplot as plt >>> x, y = np.random.multivariate_normal(mean, cov, 5000).T >>> plt.plot(x, y, 'x') >>> plt.axis('equal') >>> plt.show() Note that the covariance matrix must be positive semidefinite (a.k.a. nonnegative-definite). Otherwise, the behavior of this method is undefined and backwards compatibility is not guaranteed. References ---------- .. [1] Papoulis, A., "Probability, Random Variables, and Stochastic Processes," 3rd ed., New York: McGraw-Hill, 1991. .. [2] Duda, R. O., Hart, P. E., and Stork, D. G., "Pattern Classification," 2nd ed., New York: Wiley, 2001. Examples -------- >>> mean = (1, 2) >>> cov = [[1, 0], [0, 1]] >>> x = np.random.multivariate_normal(mean, cov, (3, 3)) >>> x.shape (3, 3, 2) The following is probably true, given that 0.6 is roughly twice the standard deviation: >>> list((x[0,0,:] - mean) < 0.6) [True, True] RandomState.multivariate_normal (line 4580) logseries(p, size=None) Draw samples from a logarithmic series distribution. Samples are drawn from a log series distribution with specified shape parameter, 0 < ``p`` < 1. Parameters ---------- loc : float scale : float > 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar where the values are all integers in [0, n]. See Also -------- scipy.stats.distributions.logser : probability density function, distribution or cumulative density function, etc. Notes ----- The probability density for the Log Series distribution is .. math:: P(k) = \frac{-p^k}{k \ln(1-p)}, where p = probability. The log series distribution is frequently used to represent species richness and occurrence, first proposed by Fisher, Corbet, and Williams in 1943 [2]. It may also be used to model the numbers of occupants seen in cars [3]. References ---------- .. [1] Buzas, Martin A.; Culver, Stephen J., Understanding regional species diversity through the log series distribution of occurrences: BIODIVERSITY RESEARCH Diversity & Distributions, Volume 5, Number 5, September 1999 , pp. 187-195(9). .. [2] Fisher, R.A,, A.S. Corbet, and C.B. Williams. 1943. The relation between the number of species and the number of individuals in a random sample of an animal population. Journal of Animal Ecology, 12:42-58. .. [3] D. J. Hand, F. Daly, D. Lunn, E. Ostrowski, A Handbook of Small Data Sets, CRC Press, 1994. .. [4] Wikipedia, "Logarithmic-distribution", http://en.wikipedia.org/wiki/Logarithmic-distribution Examples -------- Draw samples from the distribution: >>> a = .6 >>> s = np.random.logseries(a, 10000) >>> count, bins, ignored = plt.hist(s) # plot against distribution >>> def logseries(k, p): ... return -p**k/(k*log(1-p)) >>> plt.plot(bins, logseries(bins, a)*count.max()/ logseries(bins, a).max(), 'r') >>> plt.show() RandomState.logseries (line 4482) hypergeometric(ngood, nbad, nsample, size=None) Draw samples from a Hypergeometric distribution. Samples are drawn from a hypergeometric distribution with specified parameters, ngood (ways to make a good selection), nbad (ways to make a bad selection), and nsample = number of items sampled, which is less than or equal to the sum ngood + nbad. Parameters ---------- ngood : int or array_like Number of ways to make a good selection. Must be nonnegative. nbad : int or array_like Number of ways to make a bad selection. Must be nonnegative. nsample : int or array_like Number of items sampled. Must be at least 1 and at most ``ngood + nbad``. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar The values are all integers in [0, n]. See Also -------- scipy.stats.distributions.hypergeom : probability density function, distribution or cumulative density function, etc. Notes ----- The probability density for the Hypergeometric distribution is .. math:: P(x) = \frac{\binom{m}{n}\binom{N-m}{n-x}}{\binom{N}{n}}, where :math:`0 \le x \le m` and :math:`n+m-N \le x \le n` for P(x) the probability of x successes, n = ngood, m = nbad, and N = number of samples. Consider an urn with black and white marbles in it, ngood of them black and nbad are white. If you draw nsample balls without replacement, then the hypergeometric distribution describes the distribution of black balls in the drawn sample. Note that this distribution is very similar to the binomial distribution, except that in this case, samples are drawn without replacement, whereas in the Binomial case samples are drawn with replacement (or the sample space is infinite). As the sample space becomes large, this distribution approaches the binomial. References ---------- .. [1] Lentner, Marvin, "Elementary Applied Statistics", Bogden and Quigley, 1972. .. [2] Weisstein, Eric W. "Hypergeometric Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/HypergeometricDistribution.html .. [3] Wikipedia, "Hypergeometric-distribution", http://en.wikipedia.org/wiki/Hypergeometric_distribution Examples -------- Draw samples from the distribution: >>> ngood, nbad, nsamp = 100, 2, 10 # number of good, number of bad, and number of samples >>> s = np.random.hypergeometric(ngood, nbad, nsamp, 1000) >>> hist(s) # note that it is very unlikely to grab both bad items Suppose you have an urn with 15 white and 15 black marbles. If you pull 15 marbles at random, how likely is it that 12 or more of them are one color? >>> s = np.random.hypergeometric(15, 15, 15, 100000) >>> sum(s>=12)/100000. + sum(s<=3)/100000. # answer = 0.003 ... pretty unlikely! RandomState.hypergeometric (line 4361) geometric(p, size=None) Draw samples from the geometric distribution. Bernoulli trials are experiments with one of two outcomes: success or failure (an example of such an experiment is flipping a coin). The geometric distribution models the number of trials that must be run in order to achieve success. It is therefore supported on the positive integers, ``k = 1, 2, ...``. The probability mass function of the geometric distribution is .. math:: f(k) = (1 - p)^{k - 1} p where `p` is the probability of success of an individual trial. Parameters ---------- p : float The probability of success of an individual trial. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : ndarray Samples from the geometric distribution, shaped according to `size`. Examples -------- Draw ten thousand values from the geometric distribution, with the probability of an individual success equal to 0.35: >>> z = np.random.geometric(p=0.35, size=10000) How many trials succeeded after a single run? >>> (z == 1).sum() / 10000. 0.34889999999999999 #random RandomState.geometric (line 4292) zipf(a, size=None) Draw samples from a Zipf distribution. Samples are drawn from a Zipf distribution with specified parameter `a` > 1. The Zipf distribution (also known as the zeta distribution) is a continuous probability distribution that satisfies Zipf's law: the frequency of an item is inversely proportional to its rank in a frequency table. Parameters ---------- a : float > 1 Distribution parameter. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : scalar or ndarray The returned samples are greater than or equal to one. See Also -------- scipy.stats.distributions.zipf : probability density function, distribution, or cumulative density function, etc. Notes ----- The probability density for the Zipf distribution is .. math:: p(x) = \frac{x^{-a}}{\zeta(a)}, where :math:`\zeta` is the Riemann Zeta function. It is named for the American linguist George Kingsley Zipf, who noted that the frequency of any word in a sample of a language is inversely proportional to its rank in the frequency table. References ---------- .. [1] Zipf, G. K., "Selected Studies of the Principle of Relative Frequency in Language," Cambridge, MA: Harvard Univ. Press, 1932. Examples -------- Draw samples from the distribution: >>> a = 2. # parameter >>> s = np.random.zipf(a, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> import scipy.special as sps Truncate s values at 50 so plot is interesting >>> count, bins, ignored = plt.hist(s[s<50], 50, normed=True) >>> x = np.arange(1., 50.) >>> y = x**(-a)/sps.zetac(a) >>> plt.plot(x, y/max(y), linewidth=2, color='r') >>> plt.show() RandomState.zipf (line 4204) poisson(lam=1.0, size=None) Draw samples from a Poisson distribution. The Poisson distribution is the limit of the binomial distribution for large N. Parameters ---------- lam : float or sequence of float Expectation of interval, should be >= 0. A sequence of expectation intervals must be broadcastable over the requested size. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar The drawn samples, of shape *size*, if it was provided. Notes ----- The Poisson distribution .. math:: f(k; \lambda)=\frac{\lambda^k e^{-\lambda}}{k!} For events with an expected separation :math:`\lambda` the Poisson distribution :math:`f(k; \lambda)` describes the probability of :math:`k` events occurring within the observed interval :math:`\lambda`. Because the output is limited to the range of the C long type, a ValueError is raised when `lam` is within 10 sigma of the maximum representable value. References ---------- .. [1] Weisstein, Eric W. "Poisson Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/PoissonDistribution.html .. [2] Wikipedia, "Poisson distribution", http://en.wikipedia.org/wiki/Poisson_distribution Examples -------- Draw samples from the distribution: >>> import numpy as np >>> s = np.random.poisson(5, 10000) Display histogram of the sample: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 14, normed=True) >>> plt.show() Draw each 100 values for lambda 100 and 500: >>> s = np.random.poisson(lam=(100., 500.), size=(100, 2)) RandomState.poisson (line 4118) negative_binomial(n, p, size=None) Draw samples from a negative binomial distribution. Samples are drawn from a negative binomial distribution with specified parameters, `n` trials and `p` probability of success where `n` is an integer > 0 and `p` is in the interval [0, 1]. Parameters ---------- n : int Parameter, > 0. p : float Parameter, >= 0 and <=1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : int or ndarray of ints Drawn samples. Notes ----- The probability density for the negative binomial distribution is .. math:: P(N;n,p) = \binom{N+n-1}{n-1}p^{n}(1-p)^{N}, where :math:`n-1` is the number of successes, :math:`p` is the probability of success, and :math:`N+n-1` is the number of trials. The negative binomial distribution gives the probability of n-1 successes and N failures in N+n-1 trials, and success on the (N+n)th trial. If one throws a die repeatedly until the third time a "1" appears, then the probability distribution of the number of non-"1"s that appear before the third "1" is a negative binomial distribution. References ---------- .. [1] Weisstein, Eric W. "Negative Binomial Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/NegativeBinomialDistribution.html .. [2] Wikipedia, "Negative binomial distribution", http://en.wikipedia.org/wiki/Negative_binomial_distribution Examples -------- Draw samples from the distribution: A real world example. A company drills wild-cat oil exploration wells, each with an estimated probability of success of 0.1. What is the probability of having one success for each successive well, that is what is the probability of a single success after drilling 5 wells, after 6 wells, etc.? >>> s = np.random.negative_binomial(1, 0.1, 100000) >>> for i in range(1, 11): ... probability = sum(s>> n, p = 10, .5 # number of trials, probability of each trial >>> s = np.random.binomial(n, p, 1000) # result of flipping a coin 10 times, tested 1000 times. A real world example. A company drills 9 wild-cat oil exploration wells, each with an estimated probability of success of 0.1. All nine wells fail. What is the probability of that happening? Let's do 20,000 trials of the model, and count the number that generate zero positive results. >>> sum(np.random.binomial(9, 0.1, 20000) == 0)/20000. # answer = 0.38885, or 38%. RandomState.binomial (line 3909) triangular(left, mode, right, size=None) Draw samples from the triangular distribution. The triangular distribution is a continuous probability distribution with lower limit left, peak at mode, and upper limit right. Unlike the other distributions, these parameters directly define the shape of the pdf. Parameters ---------- left : scalar Lower limit. mode : scalar The value where the peak of the distribution occurs. The value should fulfill the condition ``left <= mode <= right``. right : scalar Upper limit, should be larger than `left`. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar The returned samples all lie in the interval [left, right]. Notes ----- The probability density function for the triangular distribution is .. math:: P(x;l, m, r) = \begin{cases} \frac{2(x-l)}{(r-l)(m-l)}& \text{for $l \leq x \leq m$},\\ \frac{2(r-x)}{(r-l)(r-m)}& \text{for $m \leq x \leq r$},\\ 0& \text{otherwise}. \end{cases} The triangular distribution is often used in ill-defined problems where the underlying distribution is not known, but some knowledge of the limits and mode exists. Often it is used in simulations. References ---------- .. [1] Wikipedia, "Triangular distribution" http://en.wikipedia.org/wiki/Triangular_distribution Examples -------- Draw values from the distribution and plot the histogram: >>> import matplotlib.pyplot as plt >>> h = plt.hist(np.random.triangular(-3, 0, 8, 100000), bins=200, ... normed=True) >>> plt.show() RandomState.triangular (line 3818) wald(mean, scale, size=None) Draw samples from a Wald, or inverse Gaussian, distribution. As the scale approaches infinity, the distribution becomes more like a Gaussian. Some references claim that the Wald is an inverse Gaussian with mean equal to 1, but this is by no means universal. The inverse Gaussian distribution was first studied in relationship to Brownian motion. In 1956 M.C.K. Tweedie used the name inverse Gaussian because there is an inverse relationship between the time to cover a unit distance and distance covered in unit time. Parameters ---------- mean : scalar Distribution mean, should be > 0. scale : scalar Scale parameter, should be >= 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar Drawn sample, all greater than zero. Notes ----- The probability density function for the Wald distribution is .. math:: P(x;mean,scale) = \sqrt{\frac{scale}{2\pi x^3}}e^ \frac{-scale(x-mean)^2}{2\cdotp mean^2x} As noted above the inverse Gaussian distribution first arise from attempts to model Brownian motion. It is also a competitor to the Weibull for use in reliability modeling and modeling stock returns and interest rate processes. References ---------- .. [1] Brighton Webs Ltd., Wald Distribution, http://www.brighton-webs.co.uk/distributions/wald.asp .. [2] Chhikara, Raj S., and Folks, J. Leroy, "The Inverse Gaussian Distribution: Theory : Methodology, and Applications", CRC Press, 1988. .. [3] Wikipedia, "Wald distribution" http://en.wikipedia.org/wiki/Wald_distribution Examples -------- Draw values from the distribution and plot the histogram: >>> import matplotlib.pyplot as plt >>> h = plt.hist(np.random.wald(3, 2, 100000), bins=200, normed=True) >>> plt.show() RandomState.wald (line 3733) rayleigh(scale=1.0, size=None) Draw samples from a Rayleigh distribution. The :math:`\chi` and Weibull distributions are generalizations of the Rayleigh. Parameters ---------- scale : scalar Scale, also equals the mode. Should be >= 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Notes ----- The probability density function for the Rayleigh distribution is .. math:: P(x;scale) = \frac{x}{scale^2}e^{\frac{-x^2}{2 \cdotp scale^2}} The Rayleigh distribution would arise, for example, if the East and North components of the wind velocity had identical zero-mean Gaussian distributions. Then the wind speed would have a Rayleigh distribution. References ---------- .. [1] Brighton Webs Ltd., "Rayleigh Distribution," http://www.brighton-webs.co.uk/distributions/rayleigh.asp .. [2] Wikipedia, "Rayleigh distribution" http://en.wikipedia.org/wiki/Rayleigh_distribution Examples -------- Draw values from the distribution and plot the histogram >>> values = hist(np.random.rayleigh(3, 100000), bins=200, normed=True) Wave heights tend to follow a Rayleigh distribution. If the mean wave height is 1 meter, what fraction of waves are likely to be larger than 3 meters? >>> meanvalue = 1 >>> modevalue = np.sqrt(2 / np.pi) * meanvalue >>> s = np.random.rayleigh(modevalue, 1000000) The percentage of waves larger than 3 meters is: >>> 100.*sum(s>3)/1000000. 0.087300000000000003 RandomState.rayleigh (line 3658) lognormal(mean=0.0, sigma=1.0, size=None) Draw samples from a log-normal distribution. Draw samples from a log-normal distribution with specified mean, standard deviation, and array shape. Note that the mean and standard deviation are not the values for the distribution itself, but of the underlying normal distribution it is derived from. Parameters ---------- mean : float Mean value of the underlying normal distribution sigma : float, > 0. Standard deviation of the underlying normal distribution size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or float The desired samples. An array of the same shape as `size` if given, if `size` is None a float is returned. See Also -------- scipy.stats.lognorm : probability density function, distribution, cumulative density function, etc. Notes ----- A variable `x` has a log-normal distribution if `log(x)` is normally distributed. The probability density function for the log-normal distribution is: .. math:: p(x) = \frac{1}{\sigma x \sqrt{2\pi}} e^{(-\frac{(ln(x)-\mu)^2}{2\sigma^2})} where :math:`\mu` is the mean and :math:`\sigma` is the standard deviation of the normally distributed logarithm of the variable. A log-normal distribution results if a random variable is the *product* of a large number of independent, identically-distributed variables in the same way that a normal distribution results if the variable is the *sum* of a large number of independent, identically-distributed variables. References ---------- .. [1] Limpert, E., Stahel, W. A., and Abbt, M., "Log-normal Distributions across the Sciences: Keys and Clues," BioScience, Vol. 51, No. 5, May, 2001. http://stat.ethz.ch/~stahel/lognormal/bioscience.pdf .. [2] Reiss, R.D. and Thomas, M., "Statistical Analysis of Extreme Values," Basel: Birkhauser Verlag, 2001, pp. 31-32. Examples -------- Draw samples from the distribution: >>> mu, sigma = 3., 1. # mean and standard deviation >>> s = np.random.lognormal(mu, sigma, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 100, normed=True, align='mid') >>> x = np.linspace(min(bins), max(bins), 10000) >>> pdf = (np.exp(-(np.log(x) - mu)**2 / (2 * sigma**2)) ... / (x * sigma * np.sqrt(2 * np.pi))) >>> plt.plot(x, pdf, linewidth=2, color='r') >>> plt.axis('tight') >>> plt.show() Demonstrate that taking the products of random samples from a uniform distribution can be fit well by a log-normal probability density function. >>> # Generate a thousand samples: each is the product of 100 random >>> # values, drawn from a normal distribution. >>> b = [] >>> for i in range(1000): ... a = 10. + np.random.random(100) ... b.append(np.product(a)) >>> b = np.array(b) / np.min(b) # scale values to be positive >>> count, bins, ignored = plt.hist(b, 100, normed=True, align='mid') >>> sigma = np.std(np.log(b)) >>> mu = np.mean(np.log(b)) >>> x = np.linspace(min(bins), max(bins), 10000) >>> pdf = (np.exp(-(np.log(x) - mu)**2 / (2 * sigma**2)) ... / (x * sigma * np.sqrt(2 * np.pi))) >>> plt.plot(x, pdf, color='r', linewidth=2) >>> plt.show() RandomState.lognormal (line 3533) logistic(loc=0.0, scale=1.0, size=None) Draw samples from a logistic distribution. Samples are drawn from a logistic distribution with specified parameters, loc (location or mean, also median), and scale (>0). Parameters ---------- loc : float scale : float > 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar where the values are all integers in [0, n]. See Also -------- scipy.stats.distributions.logistic : probability density function, distribution or cumulative density function, etc. Notes ----- The probability density for the Logistic distribution is .. math:: P(x) = P(x) = \frac{e^{-(x-\mu)/s}}{s(1+e^{-(x-\mu)/s})^2}, where :math:`\mu` = location and :math:`s` = scale. The Logistic distribution is used in Extreme Value problems where it can act as a mixture of Gumbel distributions, in Epidemiology, and by the World Chess Federation (FIDE) where it is used in the Elo ranking system, assuming the performance of each player is a logistically distributed random variable. References ---------- .. [1] Reiss, R.-D. and Thomas M. (2001), "Statistical Analysis of Extreme Values, from Insurance, Finance, Hydrology and Other Fields," Birkhauser Verlag, Basel, pp 132-133. .. [2] Weisstein, Eric W. "Logistic Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/LogisticDistribution.html .. [3] Wikipedia, "Logistic-distribution", http://en.wikipedia.org/wiki/Logistic_distribution Examples -------- Draw samples from the distribution: >>> loc, scale = 10, 1 >>> s = np.random.logistic(loc, scale, 10000) >>> count, bins, ignored = plt.hist(s, bins=50) # plot against distribution >>> def logist(x, loc, scale): ... return exp((loc-x)/scale)/(scale*(1+exp((loc-x)/scale))**2) >>> plt.plot(bins, logist(bins, loc, scale)*count.max()/\ ... logist(bins, loc, scale).max()) >>> plt.show() RandomState.logistic (line 3442) gumbel(loc=0.0, scale=1.0, size=None) Draw samples from a Gumbel distribution. Draw samples from a Gumbel distribution with specified location and scale. For more information on the Gumbel distribution, see Notes and References below. Parameters ---------- loc : float The location of the mode of the distribution. scale : float The scale parameter of the distribution. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar See Also -------- scipy.stats.gumbel_l scipy.stats.gumbel_r scipy.stats.genextreme weibull Notes ----- The Gumbel (or Smallest Extreme Value (SEV) or the Smallest Extreme Value Type I) distribution is one of a class of Generalized Extreme Value (GEV) distributions used in modeling extreme value problems. The Gumbel is a special case of the Extreme Value Type I distribution for maximums from distributions with "exponential-like" tails. The probability density for the Gumbel distribution is .. math:: p(x) = \frac{e^{-(x - \mu)/ \beta}}{\beta} e^{ -e^{-(x - \mu)/ \beta}}, where :math:`\mu` is the mode, a location parameter, and :math:`\beta` is the scale parameter. The Gumbel (named for German mathematician Emil Julius Gumbel) was used very early in the hydrology literature, for modeling the occurrence of flood events. It is also used for modeling maximum wind speed and rainfall rates. It is a "fat-tailed" distribution - the probability of an event in the tail of the distribution is larger than if one used a Gaussian, hence the surprisingly frequent occurrence of 100-year floods. Floods were initially modeled as a Gaussian process, which underestimated the frequency of extreme events. It is one of a class of extreme value distributions, the Generalized Extreme Value (GEV) distributions, which also includes the Weibull and Frechet. The function has a mean of :math:`\mu + 0.57721\beta` and a variance of :math:`\frac{\pi^2}{6}\beta^2`. References ---------- .. [1] Gumbel, E. J., "Statistics of Extremes," New York: Columbia University Press, 1958. .. [2] Reiss, R.-D. and Thomas, M., "Statistical Analysis of Extreme Values from Insurance, Finance, Hydrology and Other Fields," Basel: Birkhauser Verlag, 2001. Examples -------- Draw samples from the distribution: >>> mu, beta = 0, 0.1 # location and scale >>> s = np.random.gumbel(mu, beta, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 30, normed=True) >>> plt.plot(bins, (1/beta)*np.exp(-(bins - mu)/beta) ... * np.exp( -np.exp( -(bins - mu) /beta) ), ... linewidth=2, color='r') >>> plt.show() Show how an extreme value distribution can arise from a Gaussian process and compare to a Gaussian: >>> means = [] >>> maxima = [] >>> for i in range(0,1000) : ... a = np.random.normal(mu, beta, 1000) ... means.append(a.mean()) ... maxima.append(a.max()) >>> count, bins, ignored = plt.hist(maxima, 30, normed=True) >>> beta = np.std(maxima) * np.sqrt(6) / np.pi >>> mu = np.mean(maxima) - 0.57721*beta >>> plt.plot(bins, (1/beta)*np.exp(-(bins - mu)/beta) ... * np.exp(-np.exp(-(bins - mu)/beta)), ... linewidth=2, color='r') >>> plt.plot(bins, 1/(beta * np.sqrt(2 * np.pi)) ... * np.exp(-(bins - mu)**2 / (2 * beta**2)), ... linewidth=2, color='g') >>> plt.show() RandomState.gumbel (line 3313) laplace(loc=0.0, scale=1.0, size=None) Draw samples from the Laplace or double exponential distribution with specified location (or mean) and scale (decay). The Laplace distribution is similar to the Gaussian/normal distribution, but is sharper at the peak and has fatter tails. It represents the difference between two independent, identically distributed exponential random variables. Parameters ---------- loc : float, optional The position, :math:`\mu`, of the distribution peak. scale : float, optional :math:`\lambda`, the exponential decay. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or float Notes ----- It has the probability density function .. math:: f(x; \mu, \lambda) = \frac{1}{2\lambda} \exp\left(-\frac{|x - \mu|}{\lambda}\right). The first law of Laplace, from 1774, states that the frequency of an error can be expressed as an exponential function of the absolute magnitude of the error, which leads to the Laplace distribution. For many problems in economics and health sciences, this distribution seems to model the data better than the standard Gaussian distribution. References ---------- .. [1] Abramowitz, M. and Stegun, I. A. (Eds.). "Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, 9th printing," New York: Dover, 1972. .. [2] Kotz, Samuel, et. al. "The Laplace Distribution and Generalizations, " Birkhauser, 2001. .. [3] Weisstein, Eric W. "Laplace Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/LaplaceDistribution.html .. [4] Wikipedia, "Laplace Distribution", http://en.wikipedia.org/wiki/Laplace_distribution Examples -------- Draw samples from the distribution >>> loc, scale = 0., 1. >>> s = np.random.laplace(loc, scale, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 30, normed=True) >>> x = np.arange(-8., 8., .01) >>> pdf = np.exp(-abs(x-loc)/scale)/(2.*scale) >>> plt.plot(x, pdf) Plot Gaussian for comparison: >>> g = (1/(scale * np.sqrt(2 * np.pi)) * ... np.exp(-(x - loc)**2 / (2 * scale**2))) >>> plt.plot(x,g) RandomState.laplace (line 3217) power(a, size=None) Draws samples in [0, 1] from a power distribution with positive exponent a - 1. Also known as the power function distribution. Parameters ---------- a : float parameter, > 0 size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar The returned samples lie in [0, 1]. Raises ------ ValueError If a < 1. Notes ----- The probability density function is .. math:: P(x; a) = ax^{a-1}, 0 \le x \le 1, a>0. The power function distribution is just the inverse of the Pareto distribution. It may also be seen as a special case of the Beta distribution. It is used, for example, in modeling the over-reporting of insurance claims. References ---------- .. [1] Christian Kleiber, Samuel Kotz, "Statistical size distributions in economics and actuarial sciences", Wiley, 2003. .. [2] Heckert, N. A. and Filliben, James J. "NIST Handbook 148: Dataplot Reference Manual, Volume 2: Let Subcommands and Library Functions", National Institute of Standards and Technology Handbook Series, June 2003. http://www.itl.nist.gov/div898/software/dataplot/refman2/auxillar/powpdf.pdf Examples -------- Draw samples from the distribution: >>> a = 5. # shape >>> samples = 1000 >>> s = np.random.power(a, samples) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, bins=30) >>> x = np.linspace(0, 1, 100) >>> y = a*x**(a-1.) >>> normed_y = samples*np.diff(bins)[0]*y >>> plt.plot(x, normed_y) >>> plt.show() Compare the power function distribution to the inverse of the Pareto. >>> from scipy import stats >>> rvs = np.random.power(5, 1000000) >>> rvsp = np.random.pareto(5, 1000000) >>> xx = np.linspace(0,1,100) >>> powpdf = stats.powerlaw.pdf(xx,5) >>> plt.figure() >>> plt.hist(rvs, bins=50, normed=True) >>> plt.plot(xx,powpdf,'r-') >>> plt.title('np.random.power(5)') >>> plt.figure() >>> plt.hist(1./(1.+rvsp), bins=50, normed=True) >>> plt.plot(xx,powpdf,'r-') >>> plt.title('inverse of 1 + np.random.pareto(5)') >>> plt.figure() >>> plt.hist(1./(1.+rvsp), bins=50, normed=True) >>> plt.plot(xx,powpdf,'r-') >>> plt.title('inverse of stats.pareto(5)') RandomState.power (line 3106) weibull(a, size=None) Draw samples from a Weibull distribution. Draw samples from a 1-parameter Weibull distribution with the given shape parameter `a`. .. math:: X = (-ln(U))^{1/a} Here, U is drawn from the uniform distribution over (0,1]. The more common 2-parameter Weibull, including a scale parameter :math:`\lambda` is just :math:`X = \lambda(-ln(U))^{1/a}`. Parameters ---------- a : float Shape of the distribution. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray See Also -------- scipy.stats.distributions.weibull_max scipy.stats.distributions.weibull_min scipy.stats.distributions.genextreme gumbel Notes ----- The Weibull (or Type III asymptotic extreme value distribution for smallest values, SEV Type III, or Rosin-Rammler distribution) is one of a class of Generalized Extreme Value (GEV) distributions used in modeling extreme value problems. This class includes the Gumbel and Frechet distributions. The probability density for the Weibull distribution is .. math:: p(x) = \frac{a} {\lambda}(\frac{x}{\lambda})^{a-1}e^{-(x/\lambda)^a}, where :math:`a` is the shape and :math:`\lambda` the scale. The function has its peak (the mode) at :math:`\lambda(\frac{a-1}{a})^{1/a}`. When ``a = 1``, the Weibull distribution reduces to the exponential distribution. References ---------- .. [1] Waloddi Weibull, Royal Technical University, Stockholm, 1939 "A Statistical Theory Of The Strength Of Materials", Ingeniorsvetenskapsakademiens Handlingar Nr 151, 1939, Generalstabens Litografiska Anstalts Forlag, Stockholm. .. [2] Waloddi Weibull, "A Statistical Distribution Function of Wide Applicability", Journal Of Applied Mechanics ASME Paper 1951. .. [3] Wikipedia, "Weibull distribution", http://en.wikipedia.org/wiki/Weibull_distribution Examples -------- Draw samples from the distribution: >>> a = 5. # shape >>> s = np.random.weibull(a, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> x = np.arange(1,100.)/50. >>> def weib(x,n,a): ... return (a / n) * (x / n)**(a - 1) * np.exp(-(x / n)**a) >>> count, bins, ignored = plt.hist(np.random.weibull(5.,1000)) >>> x = np.arange(1,100.)/50. >>> scale = count.max()/weib(x, 1., 5.).max() >>> plt.plot(x, weib(x, 1., 5.)*scale) >>> plt.show() RandomState.weibull (line 2997) pareto(a, size=None) Draw samples from a Pareto II or Lomax distribution with specified shape. The Lomax or Pareto II distribution is a shifted Pareto distribution. The classical Pareto distribution can be obtained from the Lomax distribution by adding 1 and multiplying by the scale parameter ``m`` (see Notes). The smallest value of the Lomax distribution is zero while for the classical Pareto distribution it is ``mu``, where the standard Pareto distribution has location ``mu = 1``. Lomax can also be considered as a simplified version of the Generalized Pareto distribution (available in SciPy), with the scale set to one and the location set to zero. The Pareto distribution must be greater than zero, and is unbounded above. It is also known as the "80-20 rule". In this distribution, 80 percent of the weights are in the lowest 20 percent of the range, while the other 20 percent fill the remaining 80 percent of the range. Parameters ---------- shape : float, > 0. Shape of the distribution. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. See Also -------- scipy.stats.distributions.lomax.pdf : probability density function, distribution or cumulative density function, etc. scipy.stats.distributions.genpareto.pdf : probability density function, distribution or cumulative density function, etc. Notes ----- The probability density for the Pareto distribution is .. math:: p(x) = \frac{am^a}{x^{a+1}} where :math:`a` is the shape and :math:`m` the scale. The Pareto distribution, named after the Italian economist Vilfredo Pareto, is a power law probability distribution useful in many real world problems. Outside the field of economics it is generally referred to as the Bradford distribution. Pareto developed the distribution to describe the distribution of wealth in an economy. It has also found use in insurance, web page access statistics, oil field sizes, and many other problems, including the download frequency for projects in Sourceforge [1]_. It is one of the so-called "fat-tailed" distributions. References ---------- .. [1] Francis Hunt and Paul Johnson, On the Pareto Distribution of Sourceforge projects. .. [2] Pareto, V. (1896). Course of Political Economy. Lausanne. .. [3] Reiss, R.D., Thomas, M.(2001), Statistical Analysis of Extreme Values, Birkhauser Verlag, Basel, pp 23-30. .. [4] Wikipedia, "Pareto distribution", http://en.wikipedia.org/wiki/Pareto_distribution Examples -------- Draw samples from the distribution: >>> a, m = 3., 2. # shape and mode >>> s = (np.random.pareto(a, 1000) + 1) * m Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, _ = plt.hist(s, 100, normed=True) >>> fit = a*m**a / bins**(a+1) >>> plt.plot(bins, max(count)*fit/max(fit), linewidth=2, color='r') >>> plt.show() RandomState.pareto (line 2893) vonmises(mu, kappa, size=None) Draw samples from a von Mises distribution. Samples are drawn from a von Mises distribution with specified mode (mu) and dispersion (kappa), on the interval [-pi, pi]. The von Mises distribution (also known as the circular normal distribution) is a continuous probability distribution on the unit circle. It may be thought of as the circular analogue of the normal distribution. Parameters ---------- mu : float Mode ("center") of the distribution. kappa : float Dispersion of the distribution, has to be >=0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : scalar or ndarray The returned samples, which are in the interval [-pi, pi]. See Also -------- scipy.stats.distributions.vonmises : probability density function, distribution, or cumulative density function, etc. Notes ----- The probability density for the von Mises distribution is .. math:: p(x) = \frac{e^{\kappa cos(x-\mu)}}{2\pi I_0(\kappa)}, where :math:`\mu` is the mode and :math:`\kappa` the dispersion, and :math:`I_0(\kappa)` is the modified Bessel function of order 0. The von Mises is named for Richard Edler von Mises, who was born in Austria-Hungary, in what is now the Ukraine. He fled to the United States in 1939 and became a professor at Harvard. He worked in probability theory, aerodynamics, fluid mechanics, and philosophy of science. References ---------- .. [1] Abramowitz, M. and Stegun, I. A. (Eds.). "Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, 9th printing," New York: Dover, 1972. .. [2] von Mises, R., "Mathematical Theory of Probability and Statistics", New York: Academic Press, 1964. Examples -------- Draw samples from the distribution: >>> mu, kappa = 0.0, 4.0 # mean and dispersion >>> s = np.random.vonmises(mu, kappa, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> from scipy.special import i0 >>> plt.hist(s, 50, normed=True) >>> x = np.linspace(-np.pi, np.pi, num=51) >>> y = np.exp(kappa*np.cos(x-mu))/(2*np.pi*i0(kappa)) >>> plt.plot(x, y, linewidth=2, color='r') >>> plt.show() RandomState.vonmises (line 2795) standard_t(df, size=None) Draw samples from a standard Student's t distribution with `df` degrees of freedom. A special case of the hyperbolic distribution. As `df` gets large, the result resembles that of the standard normal distribution (`standard_normal`). Parameters ---------- df : int Degrees of freedom, should be > 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar Drawn samples. Notes ----- The probability density function for the t distribution is .. math:: P(x, df) = \frac{\Gamma(\frac{df+1}{2})}{\sqrt{\pi df} \Gamma(\frac{df}{2})}\Bigl( 1+\frac{x^2}{df} \Bigr)^{-(df+1)/2} The t test is based on an assumption that the data come from a Normal distribution. The t test provides a way to test whether the sample mean (that is the mean calculated from the data) is a good estimate of the true mean. The derivation of the t-distribution was first published in 1908 by William Gisset while working for the Guinness Brewery in Dublin. Due to proprietary issues, he had to publish under a pseudonym, and so he used the name Student. References ---------- .. [1] Dalgaard, Peter, "Introductory Statistics With R", Springer, 2002. .. [2] Wikipedia, "Student's t-distribution" http://en.wikipedia.org/wiki/Student's_t-distribution Examples -------- From Dalgaard page 83 [1]_, suppose the daily energy intake for 11 women in Kj is: >>> intake = np.array([5260., 5470, 5640, 6180, 6390, 6515, 6805, 7515, \ ... 7515, 8230, 8770]) Does their energy intake deviate systematically from the recommended value of 7725 kJ? We have 10 degrees of freedom, so is the sample mean within 95% of the recommended value? >>> s = np.random.standard_t(10, size=100000) >>> np.mean(intake) 6753.636363636364 >>> intake.std(ddof=1) 1142.1232221373727 Calculate the t statistic, setting the ddof parameter to the unbiased value so the divisor in the standard deviation will be degrees of freedom, N-1. >>> t = (np.mean(intake)-7725)/(intake.std(ddof=1)/np.sqrt(len(intake))) >>> import matplotlib.pyplot as plt >>> h = plt.hist(s, bins=100, normed=True) For a one-sided t-test, how far out in the distribution does the t statistic appear? >>> np.sum(s>> s = np.random.standard_cauchy(1000000) >>> s = s[(s>-25) & (s<25)] # truncate distribution so it plots well >>> plt.hist(s, bins=100) >>> plt.show() RandomState.standard_cauchy (line 2626) noncentral_chisquare(df, nonc, size=None) Draw samples from a noncentral chi-square distribution. The noncentral :math:`\chi^2` distribution is a generalisation of the :math:`\chi^2` distribution. Parameters ---------- df : int Degrees of freedom, should be > 0 as of Numpy 1.10, should be > 1 for earlier versions. nonc : float Non-centrality, should be non-negative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Notes ----- The probability density function for the noncentral Chi-square distribution is .. math:: P(x;df,nonc) = \sum^{\infty}_{i=0} \frac{e^{-nonc/2}(nonc/2)^{i}}{i!} \P_{Y_{df+2i}}(x), where :math:`Y_{q}` is the Chi-square with q degrees of freedom. In Delhi (2007), it is noted that the noncentral chi-square is useful in bombing and coverage problems, the probability of killing the point target given by the noncentral chi-squared distribution. References ---------- .. [1] Delhi, M.S. Holla, "On a noncentral chi-square distribution in the analysis of weapon systems effectiveness", Metrika, Volume 15, Number 1 / December, 1970. .. [2] Wikipedia, "Noncentral chi-square distribution" http://en.wikipedia.org/wiki/Noncentral_chi-square_distribution Examples -------- Draw values from the distribution and plot the histogram >>> import matplotlib.pyplot as plt >>> values = plt.hist(np.random.noncentral_chisquare(3, 20, 100000), ... bins=200, normed=True) >>> plt.show() Draw values from a noncentral chisquare with very small noncentrality, and compare to a chisquare. >>> plt.figure() >>> values = plt.hist(np.random.noncentral_chisquare(3, .0000001, 100000), ... bins=np.arange(0., 25, .1), normed=True) >>> values2 = plt.hist(np.random.chisquare(3, 100000), ... bins=np.arange(0., 25, .1), normed=True) >>> plt.plot(values[1][0:-1], values[0]-values2[0], 'ob') >>> plt.show() Demonstrate how large values of non-centrality lead to a more symmetric distribution. >>> plt.figure() >>> values = plt.hist(np.random.noncentral_chisquare(3, 20, 100000), ... bins=200, normed=True) >>> plt.show() RandomState.noncentral_chisquare (line 2529) chisquare(df, size=None) Draw samples from a chi-square distribution. When `df` independent random variables, each with standard normal distributions (mean 0, variance 1), are squared and summed, the resulting distribution is chi-square (see Notes). This distribution is often used in hypothesis testing. Parameters ---------- df : int Number of degrees of freedom. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- output : ndarray Samples drawn from the distribution, packed in a `size`-shaped array. Raises ------ ValueError When `df` <= 0 or when an inappropriate `size` (e.g. ``size=-1``) is given. Notes ----- The variable obtained by summing the squares of `df` independent, standard normally distributed random variables: .. math:: Q = \sum_{i=0}^{\mathtt{df}} X^2_i is chi-square distributed, denoted .. math:: Q \sim \chi^2_k. The probability density function of the chi-squared distribution is .. math:: p(x) = \frac{(1/2)^{k/2}}{\Gamma(k/2)} x^{k/2 - 1} e^{-x/2}, where :math:`\Gamma` is the gamma function, .. math:: \Gamma(x) = \int_0^{-\infty} t^{x - 1} e^{-t} dt. References ---------- .. [1] NIST "Engineering Statistics Handbook" http://www.itl.nist.gov/div898/handbook/eda/section3/eda3666.htm Examples -------- >>> np.random.chisquare(2,4) array([ 1.89920014, 9.00867716, 3.13710533, 5.62318272]) RandomState.chisquare (line 2448) noncentral_f(dfnum, dfden, nonc, size=None) Draw samples from the noncentral F distribution. Samples are drawn from an F distribution with specified parameters, `dfnum` (degrees of freedom in numerator) and `dfden` (degrees of freedom in denominator), where both parameters > 1. `nonc` is the non-centrality parameter. Parameters ---------- dfnum : int Parameter, should be > 1. dfden : int Parameter, should be > 1. nonc : float Parameter, should be >= 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : scalar or ndarray Drawn samples. Notes ----- When calculating the power of an experiment (power = probability of rejecting the null hypothesis when a specific alternative is true) the non-central F statistic becomes important. When the null hypothesis is true, the F statistic follows a central F distribution. When the null hypothesis is not true, then it follows a non-central F statistic. References ---------- .. [1] Weisstein, Eric W. "Noncentral F-Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/NoncentralF-Distribution.html .. [2] Wikipedia, "Noncentral F distribution", http://en.wikipedia.org/wiki/Noncentral_F-distribution Examples -------- In a study, testing for a specific alternative to the null hypothesis requires use of the Noncentral F distribution. We need to calculate the area in the tail of the distribution that exceeds the value of the F distribution for the null hypothesis. We'll plot the two probability distributions for comparison. >>> dfnum = 3 # between group deg of freedom >>> dfden = 20 # within groups degrees of freedom >>> nonc = 3.0 >>> nc_vals = np.random.noncentral_f(dfnum, dfden, nonc, 1000000) >>> NF = np.histogram(nc_vals, bins=50, normed=True) >>> c_vals = np.random.f(dfnum, dfden, 1000000) >>> F = np.histogram(c_vals, bins=50, normed=True) >>> plt.plot(F[1][1:], F[0]) >>> plt.plot(NF[1][1:], NF[0]) >>> plt.show() RandomState.noncentral_f (line 2352) f(dfnum, dfden, size=None) Draw samples from an F distribution. Samples are drawn from an F distribution with specified parameters, `dfnum` (degrees of freedom in numerator) and `dfden` (degrees of freedom in denominator), where both parameters should be greater than zero. The random variate of the F distribution (also known as the Fisher distribution) is a continuous probability distribution that arises in ANOVA tests, and is the ratio of two chi-square variates. Parameters ---------- dfnum : float Degrees of freedom in numerator. Should be greater than zero. dfden : float Degrees of freedom in denominator. Should be greater than zero. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar Samples from the Fisher distribution. See Also -------- scipy.stats.distributions.f : probability density function, distribution or cumulative density function, etc. Notes ----- The F statistic is used to compare in-group variances to between-group variances. Calculating the distribution depends on the sampling, and so it is a function of the respective degrees of freedom in the problem. The variable `dfnum` is the number of samples minus one, the between-groups degrees of freedom, while `dfden` is the within-groups degrees of freedom, the sum of the number of samples in each group minus the number of groups. References ---------- .. [1] Glantz, Stanton A. "Primer of Biostatistics.", McGraw-Hill, Fifth Edition, 2002. .. [2] Wikipedia, "F-distribution", http://en.wikipedia.org/wiki/F-distribution Examples -------- An example from Glantz[1], pp 47-40: Two groups, children of diabetics (25 people) and children from people without diabetes (25 controls). Fasting blood glucose was measured, case group had a mean value of 86.1, controls had a mean value of 82.2. Standard deviations were 2.09 and 2.49 respectively. Are these data consistent with the null hypothesis that the parents diabetic status does not affect their children's blood glucose levels? Calculating the F statistic from the data gives a value of 36.01. Draw samples from the distribution: >>> dfnum = 1. # between group degrees of freedom >>> dfden = 48. # within groups degrees of freedom >>> s = np.random.f(dfnum, dfden, 1000) The lower bound for the top 1% of the samples is : >>> sort(s)[-10] 7.61988120985 So there is about a 1% chance that the F statistic will exceed 7.62, the measured value is 36, so the null hypothesis is rejected at the 1% level. RandomState.f (line 2246) gamma(shape, scale=1.0, size=None) Draw samples from a Gamma distribution. Samples are drawn from a Gamma distribution with specified parameters, `shape` (sometimes designated "k") and `scale` (sometimes designated "theta"), where both parameters are > 0. Parameters ---------- shape : scalar > 0 The shape of the gamma distribution. scale : scalar > 0, optional The scale of the gamma distribution. Default is equal to 1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : ndarray, float Returns one sample unless `size` parameter is specified. See Also -------- scipy.stats.distributions.gamma : probability density function, distribution or cumulative density function, etc. Notes ----- The probability density for the Gamma distribution is .. math:: p(x) = x^{k-1}\frac{e^{-x/\theta}}{\theta^k\Gamma(k)}, where :math:`k` is the shape and :math:`\theta` the scale, and :math:`\Gamma` is the Gamma function. The Gamma distribution is often used to model the times to failure of electronic components, and arises naturally in processes for which the waiting times between Poisson distributed events are relevant. References ---------- .. [1] Weisstein, Eric W. "Gamma Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/GammaDistribution.html .. [2] Wikipedia, "Gamma-distribution", http://en.wikipedia.org/wiki/Gamma-distribution Examples -------- Draw samples from the distribution: >>> shape, scale = 2., 2. # mean and dispersion >>> s = np.random.gamma(shape, scale, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> import scipy.special as sps >>> count, bins, ignored = plt.hist(s, 50, normed=True) >>> y = bins**(shape-1)*(np.exp(-bins/scale) / ... (sps.gamma(shape)*scale**shape)) >>> plt.plot(bins, y, linewidth=2, color='r') >>> plt.show() RandomState.gamma (line 2152) standard_gamma(shape, size=None) Draw samples from a standard Gamma distribution. Samples are drawn from a Gamma distribution with specified parameters, shape (sometimes designated "k") and scale=1. Parameters ---------- shape : float Parameter, should be > 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar The drawn samples. See Also -------- scipy.stats.distributions.gamma : probability density function, distribution or cumulative density function, etc. Notes ----- The probability density for the Gamma distribution is .. math:: p(x) = x^{k-1}\frac{e^{-x/\theta}}{\theta^k\Gamma(k)}, where :math:`k` is the shape and :math:`\theta` the scale, and :math:`\Gamma` is the Gamma function. The Gamma distribution is often used to model the times to failure of electronic components, and arises naturally in processes for which the waiting times between Poisson distributed events are relevant. References ---------- .. [1] Weisstein, Eric W. "Gamma Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/GammaDistribution.html .. [2] Wikipedia, "Gamma-distribution", http://en.wikipedia.org/wiki/Gamma-distribution Examples -------- Draw samples from the distribution: >>> shape, scale = 2., 1. # mean and width >>> s = np.random.standard_gamma(shape, 1000000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> import scipy.special as sps >>> count, bins, ignored = plt.hist(s, 50, normed=True) >>> y = bins**(shape-1) * ((np.exp(-bins/scale))/ \ ... (sps.gamma(shape) * scale**shape)) >>> plt.plot(bins, y, linewidth=2, color='r') >>> plt.show() RandomState.standard_gamma (line 2066) standard_exponential(size=None) Draw samples from the standard exponential distribution. `standard_exponential` is identical to the exponential distribution with a scale parameter of 1. Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : float or ndarray Drawn samples. Examples -------- Output a 3x8000 array: >>> n = np.random.standard_exponential((3, 8000)) RandomState.standard_exponential (line 2035) normal(loc=0.0, scale=1.0, size=None) Draw random samples from a normal (Gaussian) distribution. The probability density function of the normal distribution, first derived by De Moivre and 200 years later by both Gauss and Laplace independently [2]_, is often called the bell curve because of its characteristic shape (see the example below). The normal distributions occurs often in nature. For example, it describes the commonly occurring distribution of samples influenced by a large number of tiny, random disturbances, each with its own unique distribution [2]_. Parameters ---------- loc : float Mean ("centre") of the distribution. scale : float Standard deviation (spread or "width") of the distribution. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. See Also -------- scipy.stats.distributions.norm : probability density function, distribution or cumulative density function, etc. Notes ----- The probability density for the Gaussian distribution is .. math:: p(x) = \frac{1}{\sqrt{ 2 \pi \sigma^2 }} e^{ - \frac{ (x - \mu)^2 } {2 \sigma^2} }, where :math:`\mu` is the mean and :math:`\sigma` the standard deviation. The square of the standard deviation, :math:`\sigma^2`, is called the variance. The function has its peak at the mean, and its "spread" increases with the standard deviation (the function reaches 0.607 times its maximum at :math:`x + \sigma` and :math:`x - \sigma` [2]_). This implies that `numpy.random.normal` is more likely to return samples lying close to the mean, rather than those far away. References ---------- .. [1] Wikipedia, "Normal distribution", http://en.wikipedia.org/wiki/Normal_distribution .. [2] P. R. Peebles Jr., "Central Limit Theorem" in "Probability, Random Variables and Random Signal Principles", 4th ed., 2001, pp. 51, 51, 125. Examples -------- Draw samples from the distribution: >>> mu, sigma = 0, 0.1 # mean and standard deviation >>> s = np.random.normal(mu, sigma, 1000) Verify the mean and the variance: >>> abs(mu - np.mean(s)) < 0.01 True >>> abs(sigma - np.std(s, ddof=1)) < 0.01 True Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 30, normed=True) >>> plt.plot(bins, 1/(sigma * np.sqrt(2 * np.pi)) * ... np.exp( - (bins - mu)**2 / (2 * sigma**2) ), ... linewidth=2, color='r') >>> plt.show() RandomState.normal (line 1812) standard_normal(size=None) Draw samples from a standard Normal distribution (mean=0, stdev=1). Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : float or ndarray Drawn samples. Examples -------- >>> s = np.random.standard_normal(8000) >>> s array([ 0.6888893 , 0.78096262, -0.89086505, ..., 0.49876311, #random -0.38672696, -0.4685006 ]) #random >>> s.shape (8000,) >>> s = np.random.standard_normal(size=(3, 4, 2)) >>> s.shape (3, 4, 2) RandomState.standard_normal (line 1779) random_integers(low, high=None, size=None) Random integers of type np.int between `low` and `high`, inclusive. Return random integers of type np.int from the "discrete uniform" distribution in the closed interval [`low`, `high`]. If `high` is None (the default), then results are from [1, `low`]. The np.int type translates to the C long type used by Python 2 for "short" integers and its precision is platform dependent. This function has been deprecated. Use randint instead. .. deprecated:: 1.11.0 Parameters ---------- low : int Lowest (signed) integer to be drawn from the distribution (unless ``high=None``, in which case this parameter is the *highest* such integer). high : int, optional If provided, the largest (signed) integer to be drawn from the distribution (see above for behavior if ``high=None``). size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : int or ndarray of ints `size`-shaped array of random integers from the appropriate distribution, or a single such random int if `size` not provided. See Also -------- random.randint : Similar to `random_integers`, only for the half-open interval [`low`, `high`), and 0 is the lowest value if `high` is omitted. Notes ----- To sample from N evenly spaced floating-point numbers between a and b, use:: a + (b - a) * (np.random.random_integers(N) - 1) / (N - 1.) Examples -------- >>> np.random.random_integers(5) 4 >>> type(np.random.random_integers(5)) >>> np.random.random_integers(5, size=(3.,2.)) array([[5, 4], [3, 3], [4, 5]]) Choose five random numbers from the set of five evenly-spaced numbers between 0 and 2.5, inclusive (*i.e.*, from the set :math:`{0, 5/8, 10/8, 15/8, 20/8}`): >>> 2.5 * (np.random.random_integers(5, size=(5,)) - 1) / 4. array([ 0.625, 1.25 , 0.625, 0.625, 2.5 ]) Roll two six sided dice 1000 times and sum the results: >>> d1 = np.random.random_integers(1, 6, 1000) >>> d2 = np.random.random_integers(1, 6, 1000) >>> dsums = d1 + d2 Display results as a histogram: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(dsums, 11, normed=True) >>> plt.show() RandomState.random_integers (line 1682) randn(d0, d1, ..., dn) Return a sample (or samples) from the "standard normal" distribution. If positive, int_like or int-convertible arguments are provided, `randn` generates an array of shape ``(d0, d1, ..., dn)``, filled with random floats sampled from a univariate "normal" (Gaussian) distribution of mean 0 and variance 1 (if any of the :math:`d_i` are floats, they are first converted to integers by truncation). A single float randomly sampled from the distribution is returned if no argument is provided. This is a convenience function. If you want an interface that takes a tuple as the first argument, use `numpy.random.standard_normal` instead. Parameters ---------- d0, d1, ..., dn : int, optional The dimensions of the returned array, should be all positive. If no argument is given a single Python float is returned. Returns ------- Z : ndarray or float A ``(d0, d1, ..., dn)``-shaped array of floating-point samples from the standard normal distribution, or a single such float if no parameters were supplied. See Also -------- random.standard_normal : Similar, but takes a tuple as its argument. Notes ----- For random samples from :math:`N(\mu, \sigma^2)`, use: ``sigma * np.random.randn(...) + mu`` Examples -------- >>> np.random.randn() 2.1923875335537315 #random Two-by-four array of samples from N(3, 6.25): >>> 2.5 * np.random.randn(2, 4) + 3 array([[-4.49401501, 4.00950034, -1.81814867, 7.29718677], #random [ 0.39924804, 4.68456316, 4.99394529, 4.84057254]]) #random RandomState.randn (line 1625) rand(d0, d1, ..., dn) Random values in a given shape. Create an array of the given shape and populate it with random samples from a uniform distribution over ``[0, 1)``. Parameters ---------- d0, d1, ..., dn : int, optional The dimensions of the returned array, should all be positive. If no argument is given a single Python float is returned. Returns ------- out : ndarray, shape ``(d0, d1, ..., dn)`` Random values. See Also -------- random Notes ----- This is a convenience function. If you want an interface that takes a shape-tuple as the first argument, refer to np.random.random_sample . Examples -------- >>> np.random.rand(3,2) array([[ 0.14022471, 0.96360618], #random [ 0.37601032, 0.25528411], #random [ 0.49313049, 0.94909878]]) #random RandomState.rand (line 1581) uniform(low=0.0, high=1.0, size=None) Draw samples from a uniform distribution. Samples are uniformly distributed over the half-open interval ``[low, high)`` (includes low, but excludes high). In other words, any value within the given interval is equally likely to be drawn by `uniform`. Parameters ---------- low : float, optional Lower boundary of the output interval. All values generated will be greater than or equal to low. The default value is 0. high : float Upper boundary of the output interval. All values generated will be less than high. The default value is 1.0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : ndarray Drawn samples, with shape `size`. See Also -------- randint : Discrete uniform distribution, yielding integers. random_integers : Discrete uniform distribution over the closed interval ``[low, high]``. random_sample : Floats uniformly distributed over ``[0, 1)``. random : Alias for `random_sample`. rand : Convenience function that accepts dimensions as input, e.g., ``rand(2,2)`` would generate a 2-by-2 array of floats, uniformly distributed over ``[0, 1)``. Notes ----- The probability density function of the uniform distribution is .. math:: p(x) = \frac{1}{b - a} anywhere within the interval ``[a, b)``, and zero elsewhere. Examples -------- Draw samples from the distribution: >>> s = np.random.uniform(-1,0,1000) All values are within the given interval: >>> np.all(s >= -1) True >>> np.all(s < 0) True Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 15, normed=True) >>> plt.plot(bins, np.ones_like(bins), linewidth=2, color='r') >>> plt.show() RandomState.uniform (line 1486) choice(a, size=None, replace=True, p=None) Generates a random sample from a given 1-D array .. versionadded:: 1.7.0 Parameters ----------- a : 1-D array-like or int If an ndarray, a random sample is generated from its elements. If an int, the random sample is generated as if a was np.arange(n) size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. replace : boolean, optional Whether the sample is with or without replacement p : 1-D array-like, optional The probabilities associated with each entry in a. If not given the sample assumes a uniform distribution over all entries in a. Returns -------- samples : 1-D ndarray, shape (size,) The generated random samples Raises ------- ValueError If a is an int and less than zero, if a or p are not 1-dimensional, if a is an array-like of size 0, if p is not a vector of probabilities, if a and p have different lengths, or if replace=False and the sample size is greater than the population size See Also --------- randint, shuffle, permutation Examples --------- Generate a uniform random sample from np.arange(5) of size 3: >>> np.random.choice(5, 3) array([0, 3, 4]) >>> #This is equivalent to np.random.randint(0,5,3) Generate a non-uniform random sample from np.arange(5) of size 3: >>> np.random.choice(5, 3, p=[0.1, 0, 0.3, 0.6, 0]) array([3, 3, 0]) Generate a uniform random sample from np.arange(5) of size 3 without replacement: >>> np.random.choice(5, 3, replace=False) array([3,1,0]) >>> #This is equivalent to np.random.permutation(np.arange(5))[:3] Generate a non-uniform random sample from np.arange(5) of size 3 without replacement: >>> np.random.choice(5, 3, replace=False, p=[0.1, 0, 0.3, 0.6, 0]) array([2, 3, 0]) Any of the above can be repeated with an arbitrary array-like instead of just integers. For instance: >>> aa_milne_arr = ['pooh', 'rabbit', 'piglet', 'Christopher'] >>> np.random.choice(aa_milne_arr, 5, p=[0.5, 0.1, 0.1, 0.3]) array(['pooh', 'pooh', 'pooh', 'Christopher', 'piglet'], dtype='|S11') RandomState.choice (line 1304) bytes(length) Return random bytes. Parameters ---------- length : int Number of random bytes. Returns ------- out : str String of length `length`. Examples -------- >>> np.random.bytes(10) ' eh\x85\x022SZ\xbf\xa4' #random RandomState.bytes (line 1275) randint(low, high=None, size=None, dtype='l') Return random integers from `low` (inclusive) to `high` (exclusive). Return random integers from the "discrete uniform" distribution of the specified dtype in the "half-open" interval [`low`, `high`). If `high` is None (the default), then results are from [0, `low`). Parameters ---------- low : int Lowest (signed) integer to be drawn from the distribution (unless ``high=None``, in which case this parameter is the *highest* such integer). high : int, optional If provided, one above the largest (signed) integer to be drawn from the distribution (see above for behavior if ``high=None``). size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. dtype : dtype, optional Desired dtype of the result. All dtypes are determined by their name, i.e., 'int64', 'int', etc, so byteorder is not available and a specific precision may have different C types depending on the platform. The default value is 'np.int'. .. versionadded:: 1.11.0 Returns ------- out : int or ndarray of ints `size`-shaped array of random integers from the appropriate distribution, or a single such random int if `size` not provided. See Also -------- random.random_integers : similar to `randint`, only for the closed interval [`low`, `high`], and 1 is the lowest value if `high` is omitted. In particular, this other one is the one to use to generate uniformly distributed discrete non-integers. Examples -------- >>> np.random.randint(2, size=10) array([1, 0, 0, 0, 1, 1, 0, 0, 1, 0]) >>> np.random.randint(1, size=10) array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0]) Generate a 2 x 4 array of ints between 0 and 4, inclusive: >>> np.random.randint(5, size=(2, 4)) array([[4, 0, 2, 1], [3, 2, 2, 0]]) RandomState.randint (line 1192) tomaxint(size=None) Random integers between 0 and ``sys.maxint``, inclusive. Return a sample of uniformly distributed random integers in the interval [0, ``sys.maxint``]. Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : ndarray Drawn samples, with shape `size`. See Also -------- randint : Uniform sampling over a given half-open interval of integers. random_integers : Uniform sampling over a given closed interval of integers. Examples -------- >>> RS = np.random.mtrand.RandomState() # need a RandomState object >>> RS.tomaxint((2,2,2)) array([[[1170048599, 1600360186], [ 739731006, 1947757578]], [[1871712945, 752307660], [1601631370, 1479324245]]]) >>> import sys >>> sys.maxint 2147483647 >>> RS.tomaxint((2,2,2)) < sys.maxint array([[[ True, True], [ True, True]], [[ True, True], [ True, True]]], dtype=bool) RandomState.tomaxint (line 1145) random_sample(size=None) Return random floats in the half-open interval [0.0, 1.0). Results are from the "continuous uniform" distribution over the stated interval. To sample :math:`Unif[a, b), b > a` multiply the output of `random_sample` by `(b-a)` and add `a`:: (b - a) * random_sample() + a Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : float or ndarray of floats Array of random floats of shape `size` (unless ``size=None``, in which case a single float is returned). Examples -------- >>> np.random.random_sample() 0.47108547995356098 >>> type(np.random.random_sample()) >>> np.random.random_sample((5,)) array([ 0.30220482, 0.86820401, 0.1654503 , 0.11659149, 0.54323428]) Three-by-two array of random numbers from [-5, 0): >>> 5 * np.random.random_sample((3, 2)) - 5 array([[-3.99149989, -0.52338984], [-2.99091858, -0.79479508], [-1.23204345, -1.75224494]]) RandomState.random_sample (line 1101)noncentral_chisquarestandard_exponential/builddir/build/BUILD/numpy-1.11.3/numpy/random/mtrand/mtrand.pyxsum(pvals[:-1]) > 1.0covariance is not positive-semidefinite.mean and cov must have same lengthcov must be 2 dimensional and squaremean must be 1 dimensionalnumpy.dualp >= 1.0p <= 0.0ngood + nbad < nsamplensample < 1nbad < 0ngood < 0p > 1.0p < 0.0a <= 1.0lam value too large.lam value too largelam < 0n <= 0p is nanp > 1p < 0n < 0left == rightmode > rightleft > modemean <= 0.0mean <= 0scale <= 0.0sigma <= 0.0sigma <= 0kappa < 0df <= 0nonc < 0dfnum <= 1dfden <= 0dfnum <= 0shape <= 0pnlfdbaTLb <= 0a <= 0scale <= 0This function is deprecated. Please call randint({low}, {high} + 1) insteadThis function is deprecated. Please call randint(1, {low} + 1) insteadRange exceeds valid boundsFewer non-zero entries in p than sizeCannot take a larger sample than population when 'replace=False'probabilities do not sum to 1probabilities are not non-negativea and p must have same sizep must be 1-dimensionala must be non-emptya must be 1-dimensionala must be greater than 0a must be 1-dimensional or an integerlow >= highhigh is out of bounds for %slow is out of bounds for %sUnsupported dtype "%s" for randintstate must be 624 longsalgorithm must be 'MT19937'Seed must be between 0 and 4294967295size is not compatible with inputs)))x)X)))))())x)X))x)X))))))))))))))))))))))))))))))))P))))))))))))))))))h)))))h))))`)))8)))8))))))))))()))))h)8))8))))))) )x)X)))x)X)))x)X)))x)X)))x)X)))x)X)))x)X)))x)X)))x)X)))) X)E(#P)`=(IH) =(!@)`0( 8) 0(*0)@&( () &(  )n)&)(Q)()( )@(")`() (')(} )`(")@' )',)')'$)' )'")')' ) ')'$)@')@n)#x)m)p)m)h)m)#`)`m)X)@m)P)&m) H)m)&@)l)8)l)0)l)()pl) )Pl)) l)#)k))k)A)`k)&)0k))j)G)j)L)gj) )`j))Yj))`')%' )Wj))')')u').')Uj))' )' x)@'p)' h)x' `)' X)o' P)f' H)@'@)~' 8)' 0) ' ()'  )s' )' )' )' )' )g')')Sj))&'):')"')y')4')')r')')Qj))')]' )').'x)'p)('h)'`)' X)k'P)'H)'@)Y'8)'0)d'()]' )Oj))')0')"')')' )')')')')h' )') ')' )Mj))')')T' )V')')' x)' p)'h)K'`)O'X)'P)'H)'@)'8) '0)'()' )')')')')')]' )')K' )')Kj))')')')')R' )')G' )B' )' )' )'x)'p)'h)'`)'X)H'P)-'H)' @) '8)Ij)0)'()' )')')')')')' )A')*')')')')9' )')Gj)):')' )')')')')'x)}'p)'h)'`)3'X)'P)='H)'@)'8)0' 0),'()' )')' )'' )')')' )x')')' )s')' )')')n')')i')d')_')')')'x)<' p)'h)'`)' X)'P)'H)Z'@)' 8)1' 0)U'()%' )')')')')') ') ' )P')K'ؿ)' п)'ȿ)')F'30P`))ж))E&h`)C1'x`' 'FSpP@ '0`'o @''5@'Z'P'}'[Pr'o'@ b'(^'W'U'K'A'x`5'@+'0#'j''m'a&S&8@`&&PW &M@&xC`&[:`&&C+&p@& v&l&(#@d& [& @U&G&0=&HpT.&0#&p &u `&``&:`>'r\'NG' 'x'@o '`'(5 'e'LUmtrand.so.debug7zXZִF!t/-W ]?Eh=ڊ2NI 5Cބ7gax/Ov YU:8 9t*ao@D]@g?uH 0ۤ,HaL]x ;z 0S[>ŝ4}?OxoSKqszҧ*yQA+nb6 !8Md*Yן*Lhp npW i_Q:Fd0KL`X8H9k@+r0b? k%2c>&le"#H̘N4cŀtGn)J#*AMhLYagXlƮ{'$y\' Sxץ|gD~ʖ@b ;*xWeiQfK/H͍֬^sWy9YxЎ.e`I=dw2ndq嫺rU p~Ÿ \Q$kT2ȚӊJ3HABU1Uz4uDNSϤbd86F@6鬺ǵVKL4-i\7Z"Q}s"QyTV, P|0&d x9'+DzE^TǨV#;j9MBA2*'t̫Ra1;Q]9VXveBJYuh*A)*tJw;]oX՚ncS$>Sok=> pWtvPHqh"<$`}[Izmr'U`/M? )wFWPB- BX U6q$Wя):ڀN5ZdbKy<4)FÛ) 9ۚ1LiYv>(zQKm!. 5vvFt%DTa5i.Y_PgWL|v\#|FM<57灜'ld5Tai$JG_@MC(!$;$( ~i.UӧMr/@gHY,`v@Yx&mJBI3qrbvNX-Jr5ۢg.E?(ճ dB|L6R6\6:e0BZ?18kaܝ*1V?ԙ*7c`ڵp,}T4TLIXGjݔ~qa2 39*bbnT KÍ5Jgӂ `)7B,Hm8 ]"X'Lv׶b;Sz؄ bzۻ { T@Lvi@(0./ O."@Q Ku T\ Mu6ep>l_Cd-:-6l&|0 HЋ]޺{_ĝzvȩ=6sJ,8̍7ɘR2}1&ϐ<*6kAݘdRr.&k%ovڗٗ+= 8"ԁ#t >0>AjRS4ZOxԮ] v>OWDFsC9K߂ 2R;ѣr{¿KnbNԸ3]>&8~.aѴDHh٠Ǝ\\90kͮnDDMY8ƀK'7v%zA≒yYs-8sv(9A=ګxsQ [7[bŸdd+eRzCoJB~SD0OHп~ݣ@[ݗLbF8k"%x#,ȘIqwRG#2H!2d]m/`ߢ?h _zUɻ.=P-G]~d5pCu`]'.Z,RgYZ.shstrtab.note.gnu.build-id.gnu.hash.dynsym.dynstr.gnu.version.gnu.version_r.rela.dyn.rela.plt.init.text.fini.rodata.eh_frame_hdr.eh_frame.init_array.fini_array.jcr.data.rel.ro.dynamic.got.got.plt.data.bss.gnu_debuglink.gnu_debugdata $o( P0((n 8o##Eo8%8%@Tx%x%f^Bxx  h@@c``@ nt`` z pp4& &  &  & ( &( 0 &0  & &&p @)0  0 D $