From 5b31e1f821de4d3483d1ca7d55b989c60aa5c965 Mon Sep 17 00:00:00 2001 From: 1f2b5799c7e6329c0238867cf93010b4 <1f2b5799c7e6329c0238867cf93010b4@app-learninglab.inria.fr> Date: Tue, 26 Sep 2023 22:38:57 +0000 Subject: [PATCH] done --- module3/exo3/correlation.png | Bin 0 -> 17400 bytes module3/exo3/exercice.ipynb | 378 ++++++++++++++--------------------- 2 files changed, 146 insertions(+), 232 deletions(-) create mode 100644 module3/exo3/correlation.png diff --git a/module3/exo3/correlation.png b/module3/exo3/correlation.png new file mode 100644 index 0000000000000000000000000000000000000000..dd0baa2a774171d41ce4bd08a431cfe041f27ae7 GIT binary patch literal 17400 zcmV)iK%&2iP)U8P*7-ZbZ>KLZ*U+lnSp_Ufq@}0xwybFAi#%#fq@|}KQEO56)-X|e7nZL z$iTqBa9P*U#mSX{G{Bl%P*lRez;J+pfx##xwK$o9f#C}S14DXwNkIt%17i#W1A|CX zc0maP17iUL1A|C*NRTrF17iyV0~1e4YDEbH0|SF|enDkXW_m`6f}y3QrGjHhep0GJ zaAk2xYHqQDXI^rCQ9*uDVo7QW0|Nup4h9AW240u^5(W3f%sd4n162kpgNVo|1qcff zJ_s=cNG>fZg9jx8g8+j9g8_pBLjXe}Lp{R+hNBE`7{wV~7)u#fFy3PlV+vxLz;uCG zm^qSpA@ds+OO_6nTdaDlt*rOhEZL^9ePa)2-_4=K(Z%tFGm-NGmm}8}ZcXk5JW@PU zd4+f<@d@)yL(o<5icqT158+-B6_LH7;i6x}CW#w~Uy-Pgl#@Irl`kzV zeL|*8R$ca%T%Wv){2zs_iiJvgN^h0dsuZZ2sQy$tsNSU!s;Q*;LF<6_B%M@UD?LHI zSNcZ`78uqV#TeU~$eS{ozBIdFzSClfs*^S+dw;4dus<{M;#|MXC)T}S9v!D zcV!QCPhBq)ZyO(X-(bH4|NMaZz==UigLj2o41F2S6d@OB6%`R(5i>J(Puzn9wnW{e zu;hl6HK{k#IWjCVGqdJqU(99Cv(K+6*i`tgSi2;vbXD1#3jNBGs$DgVwO(~o>mN4i zHPtkqZIx>)Y(Ls5-Br|mx>vQYvH$Kwn@O`L|D75??eGkZnfg$5<;Xeg_o%+-I&+-3%01W^SH2RkDT>t<81ZP1_K>z@;j(q!3lK=n! zAY({UO#lFTB>(_`g8%^e{{R4h=>PzAFaQARU;qF*m;eA5Z<1fdMgRanN=ZaPRCwC$ zon=5&>D$JyA%^Y{EIYBvcSYx@)fY z!Yg{}VORw}VdqVM#bjr!G zbkltj|Fzt!m{0D2xn(53*74aSJIirj`htL`{kEJ!eF{_^fQ3w6NRf zPwf9ERqj{u+F_YIrbFk(O+6blE3f-!KEFKJ?1S9@62e2?1%^tmYlCaU=C!JLl&tuc zOvt=LuO8<4SftKwi7RE>&ca{Pznovy zuWktgU6Dxk%>w8HI%2)jpA#lIHa?n4k|BLX*w2Imp*hcBD^Pkxx)PR>1JwYOzmhMz zM8-XPBi&2fwhk5DhRGhBB;i9!{_oG@?Sm`m13JA(GJR_xnfmtd~4! z*f(=4Nd}=eci!Yo=-nlmAbw@u+N*2%vtX?039^mv(QJlkvjvMxab1F@Az7U?xe?PGP;=4Azr8iPg6;k0gg+jV&xia9Ekl5hykanM0|0czvc)6n*I%!n#$q~EC&^bhcpCa1Ui%*&e5Y9Uxq-S_e>(+llJr5l z-snP-$1qa*V3(~&k||a3OuBwK+eLCdL|txhBu>b7a211j8(GON_dPi6oz+&1&eJAnA=26~r$t zcQZmaj+9OHB+10m`T-=_60vh}`82SHtZmTeiVGq0_8H5FoFiRCV_JeNq ztiSESDw0eo2Qdzl}%&P`@T2qu$A0~9*|=IuJ8SGS%5dCb;@y5HH{)lOApC zgB~p7$?o{H3?VK1##Q4qe${QgCoY}btX#dxdrnmmNhboB>Bd~>Afca^US5?XJP*M3pdWf7Mn?q&zjLQ<^QrxcmPvx ztsh?wBWTf&66GvoZA)ee;D>~FIjOW;J$UwLJ&(Hppe;$ZER`xe(2u0scEb%vyU!$< zE`T)cWi|tEbmBP03!;bU*5#uQo#$b2x|%eZiys&BVs;Q zb2H9nsm$g%=h*5VWTc_=ZEFGNSdwHL0J?YOw`&eS$E27}i8zI_&znX7N=?6OVe*0G zzX0ZtWOdKAShN=3^A4u8Hi#g}hIk$4-O95R)@HKYD@Zs;E&)z4B&TEg-f5S!pKs-g zyaoU;M%!dNl&Dl|xiXf?PWD;1?gGY;DG!b>%u2GfbGKquzY}(WMi|wLwa|$ z@JRB2bGh=U4zS9}K3&S}8>z`myGn;slNiL>gBpFexa^_pG zPngWTeF4!rlCuGzPmc5TB{H?SMpk<)EL1a0*9=(K*ddX8=8W4+k&t8Yqp*tU%>b{;Wn{av}b?iOyknos8cOhi%6m9qUyw@8E8?e$fDX zQKr0}YTA-o=q3_|J@`|Mmx&WgDp%i z-J2w9G@?u8TaPcU>}uSBCv{$s>3ty6b|eX(#=@fsJyoy~`3@M=%NuJf59uLB>>?LXvyWFupDT`@hJ?CIB37$#&r( zNN=--)MekjBzIT=IywMgFUK+|jSP(^$#n(~O9L?IdQrg~y9^N7H|^54&$3LRJbI8W z8by*ZzqKEHkz~r_i~stEMy4mN-*)W5yHt6hS;CBcEo+VOm*08Bs%XBI%Gcrdm=ei| zaA77qkxc&MOOihIQ9e{NKZzvMjCmrI7f1fsbNJp5xy<7WAS%^k{*Ay7afyoF|D-^9 zUr|cV-AIy^*w$kvJGo9xAbIQ1Yhj1ID(xW2_)HydMeTuGBS`Ym$o{LdTv5;0Px5Nt zdvs7hs8rDi-dx+EtY_CXPm=OHQ@`zezHjf1FoinF@=|l2;yyidIq;irpE++h$r{pXOhVg{`W3!`728L`{Y!?@)IBFFS{EN zF9qc>Ay~UC!+PqO49+W@EnV2&-{;x$$1;`Cr+olP#*Mr}ru@5M-R0XKDDz_->DFM- z6$wdZe7Le=K!<~m+ql}4nG&NO%d>VsS$UXgAA@2fBpJMN@RS?U`;+qF=B`z~3ildy zLC$8LkW=Rk8M2Q2w=w`Ue96nb6G>)_KF#FKy&6}rGj@KGv87(wPRnBzu{;2$*Am`K zOg-^qWk+3|VsB%eVen#06DKP}0C2DOiDb%Z=lbNA_)Li76Y4(Ws8W%+}^fjF2SXwi+2vlF89&(TNO=AoQ`mFte@lZx!fT0#)8AVA(}#fBq4Gj)z+-$qdxNgbuI} z88+`eYT*!EVCs&lfMGGgwrAKR`Th=)N1;8gb-%bfjQVpUnlwNzQ-~4u%O)7yD4lxqLdwKy2g$cwP4WoEY5{Lm$UI8;*t< z`LaxNwGjL{TgToRfS2S`ohq@O#a=y+Dqn#lH{;l&+CcZ450hTJ@^h*f8kNRz2w;E% z{z_jOydP+CNjhB|6QlLXZ4L*j?T|hkHR4I~dSw7sPe?MLWAV?*SWd%2;X_5iza3N3 zzAZ+uE~;%0`ncvG>%*JzOgJRiF!@Q`og`VM7{}lb;kyP1BI$<)B%_-G@OqIb+@k{| zlN-y$FDYkP558n#!V@bPp=t()VsISE^>C?)9wfOOcizJDrtq$JD=92M*U_yb0El!T zcEsZEtW!TYU>$FZ>vAD5gN!n6NRm8-(^;n7I2ZA+5egJrCPLvHvv zktAP{_7>NMDA-*AYobsL$#`t0-5=q9})Fj^yhHHJ;H0v8e$^w@HFkE$(lZ4=oyQq@H z;Ew432+c_z5>?<)EEcf6dW;&2#!Y!b(xYrTNxs7OVcA6n?y zJ|h?5DyO3Ew4Yn+mrLOq_>!E4$0*8*5bBd;Vi^F2zs0eI?x};*K9XF8 z(c;F`)TB~^q7kPgC&@=yl=;5aFSJS@CJW7GK*b@pDp^^r_;s7jJ?W#G}Gt_!{= zx8f;Dk9urn9!v6r&ZM8jiAV8UhN+>*9~ms-U${lGX#?pmezHzru)u3X>6>yv z!m$(Z7s+`L{QEDlgk(x(I9md!oWbx*(`&%?H21U;8*fs^^0E+iM=}r-NOBGgDt*7Y zyEN>31vZl;2gt-&#+%QU&dsOoT?7YLw55lHWO|v_`**w{S*w`T;W`)R$=40x^=8~n z6hGW$%DwT&W$+g8pPzl5AJ|PFDCixfDmp%tvYm z?vcC;z@g8aI?^bgn;1eepcD2fVVT-ce3m5tcC9Q>t!wd;8ljQ6=;A6H&q_uU2|^dh~ync#`!Ok7QvZ>G1HZL=b-m9Z=?JRt1_ zeb1@iF_@4$nY9P-55 zyjG&K2YFE>lD-T>b>kmk5lM2SD4e}Z3b$ks0El&A_9;w|wfGhb`R>UMQ;$#u`sGl} zA-yr0WYe;2ZZ2pj!_Csp<2N#`bPN9R8@bb0O z`|Euq8JRvyC!FLWtbGvhZlp{pYsLc^+lV z;Ch?vC=Ih$VZE$k_LTU5Tq+i<=0;_&Pi4vh>- z&g%L7j3zn(Bsm%H*#-V&vSTAA)qx1GxgV`xhvZC%?nx(@_wS{_^TY6-!=?E*i6>W$ z`=Q0y(Ovmb-23pc({~;pF|e=^0q$=U*?p4Rpg+=0!| zr3yUXll)uHpQNPH3XVa;b+g<}s>2M1!Yi^8LoToBZP0zFdB~NuzMd6?*)*e%g`hLWcz?{O1E6~{y-=krgj_3<M^ zer_xnP5L^-B-@OX>H#MhlijkHp3L~t!rALpu`1iB&Gb)-?Xd~HfijOtGUO9KO9MgU zZ}79tpCtGD@KHGwpAkip;X*8$e>~&70F=BC)2aMasX!Ct3;^zzX2TK?VJJy*jb7|O zuq=gUpINo~HaBeN@0l3#wZ}gG>qI!o17g6s#n4Qn|C=!*H^h@4o57&&>bxR&5(l&% zvH5*gfKj^iWAckzC$F5pn?|Nbsg`;x$MS(K#Qm>?@qmGj9lE9Q4u+HbTo0G;2zvC< zR*xCF#%t`VQ~v3UJb&-g+xuq)`7JS!OwGqlktU7azqglsD40D%55(oQ<!X1kH^!5Xh18%h@3&#S zU;n@NMJMKbyae8#w`>R|`RIJ|f7*YUdNAtCpWnB*g8zlMf^RJIVrllDVfj~~vrJcu zWs2+9#?tgZ!*aL!EZ?90ed;gc*Qyv9^soCpwOE$@#!?F`6Qwa9A$cV*BuoDoFs81a zpy*_VQ>WiqGh^1WV=ttR^lMoW;u*=L7mj^nS>!C+v=$`)HuXp{Fnb>7kJ7S6$DrHT zd;8`+-Cd)zd|bkqm`ZZ40j6dC$h$rRjVh+3J=-~ z7f$k}k>gb|A?WG;Y0Ee6I~)5Ujob?u&Ws@63K?eoalI&51{gzpoFp5Rm9RSi5lUWp zY9N@{(@OQo{K-YE$hO1nI8Bm~&9UJWEKc?XMsGmxbdn>D(oPQ_1-(wqF#W;5JC~CXjxAJmYvpo6PyEP# z(ShVuG~;fVOj&-3WP1y5Q|w<~bV#eULd@443aB=vl-cQPOnORil40^zCcQN0RTz{lcS1D-=_8^ zp9{qDlZj3Amt`66fhmG^6{y=hsk~n39>Fiu+$g-c)EjjN!+28H@iX}Z<=IC?8M+BoAFEwOxsmGrQQW{mOhU{pya$M-?Nq^KLc>v><3DrZ!;E34kFSOx4!LOdrSfU%DC6=E=bGZV4GZ4%7 z36_7(vssLXtXG}fgsd2yxyYVE|3Es&j)BzKKce_x94ucVf^}*LTLj~mg0W04i2cIa^^oSyO-CzMwf6NIpe7lG|{kuAVGRANcUx{n-3J$K2uF$QUp2Gb{uny-$WcegUh_YvcpBKWnZtClZd`a>ymX1e+ zH|A3sW5#RJIxw2+myO_4Fe;M#Qe%?BRblxev-r;iSPvc26EzPA;SRlZJ~B2nqB(rt z3zaGGJ$VscB-db{xo~9@^+sSbLu{QV%OZCRnPsf$7;*wm@5Rk!xJr@%;=R7O-ygWb zy3#jHl9JAL_{aCix~w~X0KO#OlqoeF4@icTZOlFfG0`s8r}VqX4syMKEQd?q{d~tm zf5tvNf9l5V7v#%4l>1{=@{d#8>022L4kO9Y&d2@cjKGC)ZwP(r$D8tB2rC239QL z$DVl2jWC5!A&65DLOys}0G3OUXHne8cVusu_= z`P2g#PEUtGGSI?i19Sn@d=TSmtq(wmoy@_~2hvhD(T?^HzhoaY2-%yPV#%4Dp@X_I0j@tcraobe7@z6G@H+sxjQ!@jg># zU^D$pP&27NP(u1J?X}j|GxuoRW!}N8oUY?==~xUU&N0N^c+q&B($C33Zi(a|l$APZ zC)OlC+Luh!oaG*&XkgCKN*eRWcOsKnK92y{$_!6* zs)0Y3q60Yu-(3ZI{N4e^dyAu(XUpboKZjoT z_qrH-8ii{lFJT0^5bH>GfL~u+BiCY4q@G6x$r-?jYseuw76!`)=tnZH5$2LhpjV?L zhLG&4f0G&8+hsnhNg*FT<%p442I!Pt97pnOIp_}zCz7h}z18&am80H25L%G~OnB*IDB$r?l z$%C_b1x{OfqCejVx=4Ioj|h$``L57e-WO5_-YtkDd8VbU$CeC|u|Z0Am~iV>kQ7GZ z<9t&vQ_jZ~fTd5SN|0^Lmf0Xy;bv#?#*?>VNcO{1(jViqZRv|0>mkCCj~K=gxI(VQ z&SYmjBg|%$xR>Ph476hvm$ZE4*bnY7wt?w_9&W-TGJ6y!0`|TnKQHSu_G}u{gmrec z4;j2%7KXG=ADBtT*Az@;aS@~>PEzPBQ%A}7KaeD8(%C|zp$f}q(mLfi#&KD6dmp{1 z?f_6nLN>67ABxi3lDjH8NYHX&?c3U2E!TxhR<2KtpB|i8Tr_K zqLwslI>C>TS=RORzt<=aIC`?~5lN=1QXOcrVYHT5ezBnn%MiT=dHvKM9hm++jmeI! z*i>&szd8%aRHqsw_u_K4Eid8^KA(|DA8Zac(l4bC;CrT4!wE7%ciIF1vr=3k1`d2V zpI5_*{9j-xD@>bnR4Bziv6m09jtzt&=ZrG!mIupoSS`KY3+JJ0T*kv4^~oa`Oom}9 zgZH^&YaJL?>g)H#Sf&*3^gqEeqpWA1zindhx(XU`4eqeIR62^#9YOLG4rbf%s5Dt& z4t!SPXiRDb8=6Nk(!x2J;X1QP`kA}i)(p*1;PWc%@hpe`>HjNBc7$JEUr#|DU2C(@ zpWv5K!hZXsfI#23F(%)We1##D#HKML)1Wa=`hzz^gk4R-HOAXJB{RyPJ4tp2DoVYi z?*<&i@9>2|3;~>9`40zH8c{xSJ%g6aZZv;B;r5bJu_VtG?N;CT zjxfgc0jTGu4~XAykrlWvQkIqL@_&Qn*r^mMLypm;%0DmS+rZeQA~_aiNG<(o3rC zYj-G)FD^QkJ%8etX89nq#bcqfJeHXhuf>tk?Znr?a`3e;&N4(U6R40`M%bEs(E2TW zlvR?kC|O3_S~sKLcR$X!km-C*mwa)S(oSwg!IJ%z6ywok?82rZXBqluQ%hO4ly&#J z1=*E$S^i`lUsNoQDGcD(Z2XG*MaTK0KO$PrlcpJKU1CdikVmih?BhJl-{gG_%2u9bPM)jVLMsC80fc;)&1A3~4!uYe^N zWEaHlnggimed@vE>u1MUS^$QNzEYNvC8{K8&axi>n6%Uf0Hdk%0d%SiMdpl$I{7HG z+obvAUjfUEP8gf@X2#V(yJ@>*0T~gW&UFDsew{35WcIAmfaQ(M*Ad3ULN2>XKPw1V zUWRIUJI0sF5&{dGr6eICDPFO2KB^U3W!9FM3HKazCe}e=X7?9k`BbM1HD}pdM$&O< z^v8ihTLf`Ol+NGw@Ev~4`w!#tO@Y%Q%ji=RTh(-OE^S+`?*Te zo8{|aLKil;cvu0|SFcfE+gLVGI;$RlW3I@&v>I6ds-Yfv?rB%@TSScDWROMh#?apT?DW#mBiY?_bVLT;?dyWU52`JPXjv&`s@ z9Al2cvxzTCs?YLN<9<`rH3 zR;>403oM@-0f4C6+ZCwwhYLypKs813Ps=RqD7YVfGfP(2yo~us3xtF$lIu=)13r^{ zr{|iYWald}CXa8u&|VArWs)0!x*=(O@xx>TjOh*h!mf&Jy$kX(nq5O{EwJH zajpeMvA^SDj_N_=a|fXZd||Roa0akB?BT-@c^An}xo(sBjO%dE9VXFUaldp`S=I+S zRsjIF4Y%t6eWFRe7v*2T{GHe)57W*Kw9o>}+7OA*Ir}X2+{{3UEoBPKvMr96OQTy# zVY8G}2kdfbuoIQr3c`zFW6tmH>fodnSjw7t0&Sb$S zR)Xo60(<{8)X!5hm|1X^&s@b~JpeHDdQFmPH;0&^OPmTfzdN8)wo{KfCIE^RICoVg zmfL{3IbWt@3MK9=NmeAUpr_7-v?+H%S-#MR2(azD(pT=@&Q_2DaV1Z6xUH;yOo7%* zP6A*g0w8WS?#DhA0nr%MR=%tKO96mwuXUE_m{JrhHv>44Yu??X%Kpa#BV|u$#I$@T0Ca^;b;U+JRe7fg56_*wAIrNy zK^p?7#&Rxz%GvLmzmapCN(b0z1ox$PLq0v&Gr$>8F8iv;GQb$QFeKP3%7ADgUpien z_CuW{!~9t$yIUG;-)|3>n|*a{rr*$@Ud=L~z{#ts+GCN!95*TN^7VPy&YA--G6rB= zPY+kM#4_0(C1P`Y=d19XHXv7URz_X?rWKaydxAqGYz^breDhfo42k6dRLdv=mNft% z+nl-@?keZ5r5)^C+|aV}kONWIpu94es`}-RxqYMa6-MvJmdUl6Zh*@tt*|6X9s*>A zrPUsn0Kn)?;j{c~4*<~3a(Mp+z0>L(B`zc?on$vCPkd38t1P9-hMZn=K&MEl>+AHvvH2b#0Z7hdO}m z9lHXS)KP(DaO(+MVOJbFGNO9jHJaEDDbIaUgXM1kn#=w1jGbcJAW7bF1N4g7nOL}0 zcT^!hNIjN9XhVICDC~xn*e0p3n)8H4vJcZ%-#hffabuul?|O}9szvhh0hxr}1ArK= ziP~W~2Ofs7nkR%(M<3__tW|z8@yk36^bE{&0a4SR)RJM9vJ5f>P|q&^=w?NuJDM1| zb9Wv=EZS!&$jvT!nPrt-e?>S7m*;1UifC(#TiRjS7qg83ShX0`r%Mw%ptx$&etMAq z`-m4eF27-We6LNGHvnMz0QD>0lYDMqJPS?eZ&)yMff_82129*+cE>ysx=Rg#9=~I1 z2WTf}Pg2#qe^vmV$aSH*-A#RTU>T%Emdjvdh!rQ4t|57ik!ZaOZOQuhvxr!B1c0+@ zA7KS-LrAh6W^cjLmCD3+6SKDdZ>C4h-@Fhi4yWA%WX zKAeKEd~IO#sWaBFsXE4b>DYogT#Di3J(zze5|*<7(<*47iK)>wmRy5+oQ%KHipljS zE(l99W7An{U?bmYA~0M5piWGIGL~!bJ-cE7|AgA@*E8y39S2|vnOXsRi-Ki_C*ZgY zCV}efl{GPoY=rq7gk9{R#DuI)KFI_CgEYJDaj?;Y=&sgSw!{{8#e4=B8NXF`|!H>R3YpWgb5@)%0R6#>if0KR8cXi$7FxjMQH#Zbl;!y}$SQ*E(a z005e`HaLL_#{pDK)Dp{+m_h!4N#uFdQ}0wl9k7U>t@Hy(1{xQnGPcSrI|CTZUU^0@ zD;qbA07D-Jimk%=TGup4J1pac3J)5&CVs^je>8MeuWy3_mZNZ+9ElNRyi48*TnfaA zl2LpJmn4#Gthktic37sA2QY@qp;5uR-9T5;2bD9JQW3kg!}2srxuKMTTq*??gQ__C z22;J2SmRY@8E)B-@C3i~Wbmu9EKGPb@8$wRpITX{#*Xo6*n1Kok$t3d@6-Npd5;CwU*m)hSE57M6kJ&(eU4 z%j*4DGnV5}xi;o7%odusNEq{X9Jw29NM2B^VXDkB%>$3x;a&^X;*PS7kkVMpbJ)kW z*sc|p196Dt8I(&U*;_5^O#^f$Uztkt&0L zZLquxr=xqrQ)QM*Fqk#rmsZ*=SZy7t)O&I{?(mdaMzefb zUdIn48DWFFB-3lD@wWFzwKVcO$fWz#d8P7cm!&sqvNh(B|3XbMO>XoS$|_7&SSej) zmbKx_I(Ws!>ehXGV}}hS$KWx8#g!ylV7XpMo=>kNg!1n~Syh!@Up=fL6Uzc-0aCfd zB&d*C4#SV^gV|(~o91;c8e^+qb*Pe2aF$Qult@q9V}PNBTF#rKhUiL?eGoua$6GD1 z942UmK03k#8R;Z#m8yJnfqM)YYz{~(nuObF?kiD|Wj!q7FbpL*9Gx^gRBA;*ObNGF z^lGchasz%KJ@Jfet2{|be*6O0w<4X&nM}Iu-^&Ld63isN@PP zon`5IeHEv{6`bXB1Ki{oj3RjyCK`6DyCw6w$llJPz&CTyjQADc#U3tG>VCEGvm^|EkTZbq>Pj>Smf ztLQ~Cq?+X|7(6d1%jM`uvH|Xr+z4k? z??XgobP*QoWLMlF*#@@?&QdxC!h(cwX{5TpF|LqYqBP}Ib(YP{!ui_3QGLT$Jsf7u zg5|fX%cei7tUL- z2rUAZ$?lj${tlpA0?C-&8uT*e8!?_9M~P@<+o{H~5xhx;S_m2|m%~AoK$*}VGP6h} z)qu)rBzLPGnV^K_T_G5%0UoP)Yc`+>8CP0Sp(*XXmp=@TAxEHc@dB*Rdr*guptzBp z$}Atj(2wL5plTXPrnXfx+=Id9fWa%0DXXt(iRDJ2p3m>E=9-9RZNEPuuuk_<7arrxGu5w&Kd}&->gN7O(s1Y8L&6S6ksKQcGOIRNV8;r-#nw`!H^k1iL+?NuT zt;}PkrV#bQRvtq4Imv4<@-HHm&voE`Me}sePv$O3IKB-2!WvDzhAmAIYCoDg=|a zX1A8)KPaE71(q2B=_L6I^En&qG&!9cFn!Pdx#~O@gypr&-c3W*%A>pw>6-tG#tnni zW!V@T$^Ebl*6cV9@sQ*rh@NPHrF87W1)fwpTZ<$)8{d;l)l8;U%<}il{&;893U%;? zyO0ij08Vj5!ZOgv^c@*sgcBMa#$S+{KHDly99Dzn1OzhJsH7%Egrum68ik{h{iW`M4obuRRc9G( z(TZeuXkh8T3#%|vQUj;7#xhX1VrI0&5e!v%rxUWu`fP$zY!R?buPpR(JPNf_U!>}= zAa4Guv{9$(EFWSO$#FQW{`ssD^3^6NY4)uS%d>I|cc?{4RkfYy5?PspLFO8yx2n(b zxwOgJTVsvp^jq*qm)f~%YK`SyTqAi%S zDmQ!mS_CXN$YcrV0USw}$CS+=%`Cz2FWbnBg5MgFwUa(>qL z^)+aLJ1MJ}^BFjVYlr0mXu z0hh=t+*B#Y=P^vOU*88ao3O%Vc>)emS(BvNhL%0@*R>xd zW3|EZOr~DBMqSgFS5_3q8HKqms3xt?e7Wz;^} zNco8)yqMykAyrv=XBI_z3|qBAQd7%kElLR`EhDtSa#JU+!^ajueUC)H^% zrUA<|5Bb~QNl`PqeC&VnQ9j=(Hp!~WvZ;w|sz|SzM^$_U;~C^)5qhanxI+QUVY#AW zpBktT=r%ahPbM2Nr8RS|tIG1FNIuuW8)eiBt$3xM%|0+&vCCCOmLZ09V?Pbl}MZ%{lktYhs^0#bevb(@NwGMFrrDC%!bP9{UsmO9a z=4O8Og_SyC!VTpW(tp)xW~!j4MOo`y0^tKarS-@YoO4GYJkrsBScojMtCZ2jr2)pO ziyGPdWFM^+dNx#K*JI4`M?A@4ZJYVO!*Zr+ z4$sKZHfmf?l3Om5=O#tMlS(W%oco)oTncqdPQyxCWmN+@{vWU`+MyE5cl)z9+LpXhIlDJ^yw(I&tZm;| z78c6_jQ+UEJS-NVSWxL#Vg@!5?t4^%|nEz*3{{5A)Om)*b%fsDrIS4nx zI`A9I{~*h}#+cf{-&p=XVVR+tKl&TX{{TxhYz5y~eq;HK&l9O;=1^X vhnpPBebD>Trz1jjIhIu+`igFqp7{R&B9`h1^|ElR00000NkvXXu0mjfpWl&= literal 0 HcmV?d00001 diff --git a/module3/exo3/exercice.ipynb b/module3/exo3/exercice.ipynb index 111dff9..11bb722 100644 --- a/module3/exo3/exercice.ipynb +++ b/module3/exo3/exercice.ipynb @@ -4,9 +4,55 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "# Simpson's paradox\n", + "# Subject 6: Around Simpson's Paradox\n", "\n", - "![SegmentLocal](simpson.gif \"segment\")" + "![SegmentLocal](simpson.gif \"segment\")\n", + "(Copyright: [tenor.com](https://tenor.com/search/homer-thinking-gifs))\n", + "\n", + " __Prerequisites__ : Averaging and ratio calculation, simple graphical presentation techniques, possibly logistic regression\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "In 1972-1974, in Whickham, a town in the north-east of England,\n", + "located approximately 6.5 kilometres south-west of Newcastle upon Tyne,\n", + "a survey of one-sixth of the electorate was conducted in order to inform\n", + "work on thyroid and heart disease (Tunbridge and\n", + "al. 1977). A continuation of this study was carried out twenty years later.\n", + "(Vanderpump et al. 1995). Some of the results were related to\n", + "smoking and whether individuals were still alive at the time of the\n", + "second study. For the purpose of simplicity, we will restrict the data to women and among these to the 1314 that were categorized as \"smoking currently\" or \"never smoked\". There were relatively few\n", + "women in the initial survey who smoked but have since quit\n", + "(162) and very few for which information was not available\n", + "(18). Survival at 20 years was determined for all women of the first survey.\n", + "\n", + "All these data are available in this [file\n", + "CSV](https://gitlab.inria.fr/learninglab/mooc-rr/mooc-rr-ressources/blob/master/module3/Practical_session/Subject6_smoking.csv). You will find on each line if the\n", + "person smokes or not, whether alive or dead at the time of the\n", + "second study, and his age at the time of the first survey. \n", + "\n", + "This exercise can be done in either R or Python.\n", + "\n", + "__Your mission, should you choose to accept it:__\n", + "\n", + "1. Tabulate the total number of women alive and dead over the period according to their smoking habits. Calculate in each group (smoking/non-smoking) the mortality rate (the ratio of the number of women who died in a group to the total number of women in that group). You can graph these data and calculate confidence intervals if you wish. Why is this result surprising?\n", + "2. Go back to question 1 (numbers and mortality rates) and add a new category related to the age group. For example, the following classes will be considered: 18-34 years, 34-54 years, 55-64 years, over 65 years.\n", + "\n", + " Why is this result surprising? Can you explain this paradox? Similarly, you may wish to provide a graphical representation of the data to support your explanations.\n", + "3. In order to avoid a bias induced by arbitrary and non-regular age groupings, it is possible to try to perform a logistic regression. If we introduce a `Death` variable of `1` or `0` to indicate whether the individual died during the 20-year period, we can study the `Death ~ Age` model to study the probability of death as a function of age according to whether one considers the group of smokers or non-smokers. Do these regressions allow you to conclude or not on the harmfulness of smoking? You will be able to propose a graphical representation of these regressions (without omitting the regions of confidence).\n", + "4. Submit your work in FUN" ] }, { @@ -18,7 +64,7 @@ }, { "cell_type": "code", - "execution_count": 25, + "execution_count": 1, "metadata": {}, "outputs": [], "source": [ @@ -156,7 +202,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "Check that \"Smoker\" and \"Satus\" have only two possible values" + "Check that \"Smoker\" and \"Satus\" have only two possible values:" ] }, { @@ -282,7 +328,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "Check for missing values in \"Age\"." + "Check for missing values in \"Age\":" ] }, { @@ -313,7 +359,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "First, group by \"Smoker\"" + "First, group by $Smoker$:" ] }, { @@ -329,7 +375,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "Number of deaths" + "Number of deaths:" ] }, { @@ -398,7 +444,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "Total number per class (smokers vs. non-smokers)" + "Total number per class (smokers vs. non-smokers):" ] }, { @@ -467,7 +513,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "Mortality rate (ratio of dead / number within the class)" + "Mortality rate (ratio of dead / number within the class):" ] }, { @@ -528,12 +574,20 @@ }, { "cell_type": "code", - "execution_count": 49, + "execution_count": 11, "metadata": {}, "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "/opt/conda/lib/python3.6/site-packages/scipy/stats/stats.py:1713: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.\n", + " return np.add.reduce(sorted[indexer] * weights, axis=axis) / sumval\n" + ] + }, { "data": { - "image/png": "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\n", + "image/png": "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\n", "text/plain": [ "
" ] @@ -549,7 +603,10 @@ " sns.barplot(data=df, x=\"Smoker\", y=\"Dead\", ax=ax, ci=95, palette=(\"blue\", \"red\")) # Note: in recent versions of seaborn, we should rather use errorbar=('ci', 95).\n", "\n", " ax.yaxis.set_major_formatter(FuncFormatter(lambda y, _: '{:.0%}'.format(y)))\n", - "\n", + " ax.tick_params(axis='both', which='major', labelsize=15)\n", + " ax.xaxis.get_label().set_fontsize(15)\n", + " ax.yaxis.get_label().set_fontsize(15)\n", + " \n", " plt.tight_layout()\n", " plt.show()" ] @@ -565,7 +622,7 @@ }, { "cell_type": "code", - "execution_count": 40, + "execution_count": 12, "metadata": {}, "outputs": [ { @@ -574,7 +631,7 @@ "Ttest_indResult(statistic=-3.057733462432345, pvalue=0.002276255348902961)" ] }, - "execution_count": 40, + "execution_count": 12, "metadata": {}, "output_type": "execute_result" } @@ -607,12 +664,12 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "First, let us split the data frame in four age classes: 18-34, 35-54, 55-64, 65+. To this end, let us use a function." + "First, let us split the data frame in four age classes: 18-34, 35-54, 55-64, 65+. We encapsulate this in a function." ] }, { "cell_type": "code", - "execution_count": 83, + "execution_count": 13, "metadata": {}, "outputs": [], "source": [ @@ -637,12 +694,12 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "Check that nobody is younger than 18" + "Check that nobody is younger than 18." ] }, { "cell_type": "code", - "execution_count": 84, + "execution_count": 14, "metadata": {}, "outputs": [ { @@ -661,12 +718,12 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "Display the mean mortality rates split by age class" + "Display the mean mortality rates split by age class." ] }, { "cell_type": "code", - "execution_count": 85, + "execution_count": 15, "metadata": {}, "outputs": [ { @@ -750,12 +807,12 @@ }, { "cell_type": "code", - "execution_count": 88, + "execution_count": 16, "metadata": {}, "outputs": [ { "data": { - "image/png": "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\n", + "image/png": "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\n", "text/plain": [ "
" ] @@ -771,7 +828,10 @@ " sns.barplot(data=df, x=\"Age_class\", y=\"Dead\", hue=\"Smoker\", ax=ax, ci=95, palette=(\"blue\", \"red\"), order=tuple(age_classes)) # Note: in recent versions of seaborn, we should rather use errorbar=('ci', 95).\n", "\n", " ax.yaxis.set_major_formatter(FuncFormatter(lambda y, _: '{:.0%}'.format(y)))\n", - "\n", + " ax.tick_params(axis='both', which='major', labelsize=15)\n", + " ax.xaxis.get_label().set_fontsize(15)\n", + " ax.yaxis.get_label().set_fontsize(15)\n", + " \n", " plt.tight_layout()\n", " plt.show()" ] @@ -780,12 +840,12 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "Redo the statistical tests" + "We redo the statistical tests." ] }, { "cell_type": "code", - "execution_count": 58, + "execution_count": 17, "metadata": {}, "outputs": [ { @@ -793,8 +853,8 @@ "output_type": "stream", "text": [ "Ttest_indResult(statistic=0.013780528783028018, pvalue=0.9890122465905076)\n", - "Ttest_indResult(statistic=1.783075769977916, pvalue=0.07588466008529735)\n", "Ttest_indResult(statistic=2.4009857289248324, pvalue=0.01677328896330688)\n", + "Ttest_indResult(statistic=1.783075769977916, pvalue=0.07588466008529735)\n", "Ttest_indResult(statistic=0.03927297913991199, pvalue=0.9687782291030265)\n" ] } @@ -831,12 +891,12 @@ }, { "cell_type": "code", - "execution_count": 89, + "execution_count": 18, "metadata": {}, "outputs": [ { "data": { - "image/png": "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\n", + "image/png": "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\n", "text/plain": [ "
" ] @@ -852,7 +912,10 @@ " sns.barplot(data=df, x=\"Age_class\", y=\"Smoker\", ax=ax, ci=95, order=tuple(age_classes)) # Note: in recent versions of seaborn, we should rather use errorbar=('ci', 95).\n", "\n", " ax.yaxis.set_major_formatter(FuncFormatter(lambda y, _: '{:.0%}'.format(y)))\n", - "\n", + " ax.tick_params(axis='both', which='major', labelsize=15)\n", + " ax.xaxis.get_label().set_fontsize(15)\n", + " ax.yaxis.get_label().set_fontsize(15)\n", + " \n", " plt.tight_layout()\n", " plt.show()" ] @@ -861,7 +924,13 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "We see that smokers are underrepresented among the elderly compared to the other three age classes (probably because many long-time smokers quit when they grow old and start developing tobacco-related health conditions). Therefore, in the overall analysis of question 1, $Age$ acted as a proxy for the absence of smoking habit. Since the mean mortality rate is higher among the elderly (irrespective of the cause -- after all, they may simply die of old age), this creates a spurious negative correlation between $Smoker$ and $Dead$ (non-smokers are more likely older, and thus more likely to have a high mortality rate). " + "We see that smokers are underrepresented among the elderly compared to the other three age classes (probably because many long-time smokers quit when they grow old and start developing tobacco-related health conditions). Therefore, in the overall analysis of question 1, $Age$ acted as a proxy for the absence of smoking habit. Since the mean mortality rate is higher among the elderly (irrespective of the cause -- after all, they may simply die of old age), this creates a spurious negative correlation between $Smoker$ and $Dead$ (non-smokers are more likely older, and thus more likely to have a high mortality rate).\n", + "\n", + "In other words, what was detected in question 1 was a *correlation*, but not the underlying mechanism of *causality*.\n", + "\n", + "![SegmentLocal](correlation.png \"segment\")\n", + "\n", + "(Copyright: [xkcd](https://xkcd.com/552/))" ] }, { @@ -875,12 +944,12 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "We perform a logistic regression for the model $Dead \\sim Age$" + "We perform a logistic regression for the model $Dead = \\beta \\times Age + \\alpha$ ($\\beta$ is the regression coefficient, $\\alpha$ is the intercept). Compared to the previous question, $Age$ is treated as a continuous covariate rather than categorised into age classes." ] }, { "cell_type": "code", - "execution_count": 70, + "execution_count": 19, "metadata": {}, "outputs": [ { @@ -888,238 +957,83 @@ "output_type": "stream", "text": [ "Optimization terminated successfully.\n", - " Current function value: 0.629619\n", - " Iterations 5\n", - "Optimization terminated successfully.\n", - " Current function value: 0.633082\n", - " Iterations 4\n", + " Current function value: 0.412727\n", + " Iterations 7\n", "Optimization terminated successfully.\n", - " Current function value: 0.687904\n", - " Iterations 3\n" + " Current function value: 0.354560\n", + " Iterations 7\n" ] } ], "source": [ "import statsmodels.api as sm\n", "\n", - "log_reg = sm.Logit(df[\"Dead\"], df[[\"Age\", \"Smoker\"]].astype(int)).fit()\n", - "\n", - "log_reg_smoker = sm.Logit(df_smoker[\"Dead\"], df_smoker[\"Age\"]).fit()\n", - "log_reg_nonsmoker = sm.Logit(df_nonsmoker[\"Dead\"], df_nonsmoker[\"Age\"]).fit()" + "log_reg_smoker = sm.Logit(df_smoker[\"Dead\"], sm.add_constant(df_smoker[\"Age\"])).fit()\n", + "log_reg_nonsmoker = sm.Logit(df_nonsmoker[\"Dead\"], sm.add_constant(df_nonsmoker[\"Age\"])).fit()" ] }, { - "cell_type": "code", - "execution_count": 71, + "cell_type": "markdown", "metadata": {}, - "outputs": [ - { - "data": { - "text/html": [ - "\n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "
Logit Regression Results
Dep. Variable: Dead No. Observations: 1314
Model: Logit Df Residuals: 1312
Method: MLE Df Model: 1
Date: Tue, 26 Sep 2023 Pseudo R-squ.: -0.06045
Time: 21:06:14 Log-Likelihood: -827.32
converged: True LL-Null: -780.16
LLR p-value: 1.000
\n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "
coef std err z P>|z| [0.025 0.975]
Age 0.0002 0.001 0.112 0.911 -0.002 0.003
Smoker -1.1657 0.114 -10.253 0.000 -1.389 -0.943
" - ], - "text/plain": [ - "\n", - "\"\"\"\n", - " Logit Regression Results \n", - "==============================================================================\n", - "Dep. Variable: Dead No. Observations: 1314\n", - "Model: Logit Df Residuals: 1312\n", - "Method: MLE Df Model: 1\n", - "Date: Tue, 26 Sep 2023 Pseudo R-squ.: -0.06045\n", - "Time: 21:06:14 Log-Likelihood: -827.32\n", - "converged: True LL-Null: -780.16\n", - " LLR p-value: 1.000\n", - "==============================================================================\n", - " coef std err z P>|z| [0.025 0.975]\n", - "------------------------------------------------------------------------------\n", - "Age 0.0002 0.001 0.112 0.911 -0.002 0.003\n", - "Smoker -1.1657 0.114 -10.253 0.000 -1.389 -0.943\n", - "==============================================================================\n", - "\"\"\"" - ] - }, - "execution_count": 71, - "metadata": {}, - "output_type": "execute_result" - } - ], "source": [ - "log_reg.summary()" + "We plot the resulting hazard curves (mortality rates as a function of age). \n", + "\n", + "The confidence bands are obtained using the the classical transformation of hazard rates: use the Gaussian 95% confidence interval $[mean \\pm 1.96 \\times standard\\ error]$ in logit space, then transform back into a probability (using the *expit* function)." ] }, { "cell_type": "code", - "execution_count": 72, + "execution_count": 20, "metadata": {}, "outputs": [ { "data": { - "text/html": [ - "\n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "
Logit Regression Results
Dep. Variable: Dead No. Observations: 582
Model: Logit Df Residuals: 581
Method: MLE Df Model: 0
Date: Tue, 26 Sep 2023 Pseudo R-squ.: -0.1516
Time: 21:06:17 Log-Likelihood: -368.45
converged: True LL-Null: -319.94
LLR p-value: nan
\n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "
coef std err z P>|z| [0.025 0.975]
Age -0.0154 0.002 -7.982 0.000 -0.019 -0.012
" - ], + "image/png": "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\n", "text/plain": [ - "\n", - "\"\"\"\n", - " Logit Regression Results \n", - "==============================================================================\n", - "Dep. Variable: Dead No. Observations: 582\n", - "Model: Logit Df Residuals: 581\n", - "Method: MLE Df Model: 0\n", - "Date: Tue, 26 Sep 2023 Pseudo R-squ.: -0.1516\n", - "Time: 21:06:17 Log-Likelihood: -368.45\n", - "converged: True LL-Null: -319.94\n", - " LLR p-value: nan\n", - "==============================================================================\n", - " coef std err z P>|z| [0.025 0.975]\n", - "------------------------------------------------------------------------------\n", - "Age -0.0154 0.002 -7.982 0.000 -0.019 -0.012\n", - "==============================================================================\n", - "\"\"\"" + "
" ] }, - "execution_count": 72, "metadata": {}, - "output_type": "execute_result" + "output_type": "display_data" } ], "source": [ - "log_reg_smoker.summary()" + "from scipy.special import expit, logit\n", + "\n", + "ages = np.linspace(df[\"Age\"].min(), df[\"Age\"].max(), 100)\n", + "\n", + "pred_smoker = log_reg_smoker.predict(sm.add_constant(ages))\n", + "pred_nonsmoker = log_reg_nonsmoker.predict(sm.add_constant(ages))\n", + "\n", + "se_smoker = np.sqrt(np.array([x @ log_reg_smoker.cov_params() @ x for x in sm.add_constant(ages)]))\n", + "se_nonsmoker = np.sqrt(np.array([x @ log_reg_nonsmoker.cov_params() @ x for x in sm.add_constant(ages)]))\n", + "\n", + "\n", + "with sns.axes_style('darkgrid'):\n", + " fig, ax = plt.subplots(figsize=(7, 5), nrows=1, ncols=1)\n", + "\n", + " ax.plot(ages, pred_smoker, color=\"red\", label=\"Smoker\")\n", + " ax.fill_between(ages, y1=expit(logit(pred_smoker) - 1.96 * se_smoker), y2=expit(logit(pred_smoker) + 1.96 * se_smoker), alpha=0.1, color=\"red\")\n", + " \n", + " ax.plot(ages, pred_nonsmoker, color=\"blue\", label=\"Non-smoker\")\n", + " ax.fill_between(ages, y1=expit(logit(pred_nonsmoker) - 1.96 * se_nonsmoker), y2=expit(logit(pred_nonsmoker) + 1.96 * se_nonsmoker), alpha=0.1, color=\"blue\")\n", + " \n", + " ax.yaxis.set_major_formatter(FuncFormatter(lambda y, _: '{:.0%}'.format(y)))\n", + "\n", + " ax.set_xlabel(\"Age\", fontsize=15)\n", + " ax.set_ylabel(\"Predicted mortality rate\", fontsize=15)\n", + " ax.legend(loc='upper left', prop={'size': 15})\n", + " ax.tick_params(axis='both', which='major', labelsize=15)\n", + "\n", + " plt.tight_layout()\n", + " plt.show()" ] }, { - "cell_type": "code", - "execution_count": 73, + "cell_type": "markdown", "metadata": {}, - "outputs": [ - { - "data": { - "text/html": [ - "\n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "
Logit Regression Results
Dep. Variable: Dead No. Observations: 732
Model: Logit Df Residuals: 731
Method: MLE Df Model: 0
Date: Tue, 26 Sep 2023 Pseudo R-squ.: -0.1052
Time: 21:06:17 Log-Likelihood: -503.55
converged: True LL-Null: -455.62
LLR p-value: nan
\n", - "\n", - "\n", - " \n", - "\n", - "\n", - " \n", - "\n", - "
coef std err z P>|z| [0.025 0.975]
Age -0.0038 0.001 -2.759 0.006 -0.007 -0.001
" - ], - "text/plain": [ - "\n", - "\"\"\"\n", - " Logit Regression Results \n", - "==============================================================================\n", - "Dep. Variable: Dead No. Observations: 732\n", - "Model: Logit Df Residuals: 731\n", - "Method: MLE Df Model: 0\n", - "Date: Tue, 26 Sep 2023 Pseudo R-squ.: -0.1052\n", - "Time: 21:06:17 Log-Likelihood: -503.55\n", - "converged: True LL-Null: -455.62\n", - " LLR p-value: nan\n", - "==============================================================================\n", - " coef std err z P>|z| [0.025 0.975]\n", - "------------------------------------------------------------------------------\n", - "Age -0.0038 0.001 -2.759 0.006 -0.007 -0.001\n", - "==============================================================================\n", - "\"\"\"" - ] - }, - "execution_count": 73, - "metadata": {}, - "output_type": "execute_result" - } - ], "source": [ - "log_reg_nonsmoker.summary()" + "**Conclusion**: the logistic regression confirms the analysis of question 2, namely that middle-aged people exhibit a higher mortality rate when smoking, while the young and the elderly have similar mortality rates." ] }, { -- 2.18.1