Do AIC and ROC in the above sample give consistent results? Why or why not?

## ROC Curve and calculating the area under the curve$($AUC$)$

$>$ library$($ROCR$)$

$>$ predictions predict$($model$,$ newdata$=$test, type$=$"response"$)$

$>$ ROCRpred prediction$($predictions$,$ test$Survived$)$

$>$ ROCRperf performance$($ROCRpred$,$ measure $="$tpr$",$ x$.$measure $="$fpr$")$

$>$ plot$($ROCRperf$,$ colorize $=$ TRUE, text.adj $=$ c$(-0\mathrm{.}2,1\mathrm{.}7),$ print.cutoffs.at $=$ seq$(0,1,0\mathrm{.}1))$

$>$

$>$ auc performance$($ROCRpred$,$ measure $=$ "auc"$)$

$>$ auc auc@y$.$values$[[1]]$

$>$ auc

$[1]0\mathrm{.}8714211>$ model glm$($Survived $,$ family$=$binomial$(1$ink$=$'logit'$),$ data$=$train$)$

$>$ summary $($model$)$

Ca$17$:

g$1$m$($formu$1$a $=$ Survived $.,$ fami$1$y $=$ binomia$1(1$ink $="1$ogit"$),$

data $=$ train$)$

Coefficients: $(3$ not defined because of singularities$)$

Signif; codes:

n to be $1)$

Nu$11$ deviance: $891\mathrm{.}99$ on $666$ degrees of freedom Resi$$ual deviance: $605\mathrm{.}78$ on $657$ degrees

of freedom

AIC: $625\mathrm{.}78$

Number of Fisher Scoring iterations: $5.$

## ROC Curve and calculating the area under the curve$($AUC$)$

$>$ library$($ROCR$)>$ ROCRpred prediction$($predictions$,$ test$Survived$)$

$>$ ROCRperf performance$($ROCRpred$,$ measure $="$tpr$",$ x$.$ measure $="$fpr$")$

$>$ plot$($ROCRperf$,$ colorize $=$ TRUE, text.adj $=c(-0\mathrm{.}2,1\mathrm{.}7),$

print.cutoffs.at $=$seq$(0,1,0\mathrm{.}1)$

$>$ auc performance$($ROCRpred$,$ measure $=$ "auc"$)$

$>$ auc auc@y$.$ values $[[1]]$

$>$ auc

$[1]0\mathrm{.}8714211$

False positive rate