Code for k$-$NN

library$($class$)$

library$($readxl$)$

library$($pROC$)$

library$($ROCR$)$

data $<-$ read$\_$excel$("$C:$/$Users$/$hemun$/$Downloads$/$RidingMowerskNN$.$xlsx$")$

print$($head$($data$))$

print$($table$($data$Validation$))$

training$\_$data $<-$ data$[$data$Validation $==0,]$

validation$\_$data $<-$ data$[$data$Validation $==1,]$

print$($nrow$($training$\_$data$))$

print$($nrow$($validation$\_$data$))$

train$\_$features $<-$ scale$($training$\_$data$[,$ c$("$Income$",$ "Lot Size"$)])$

validation$\_$features $<-$ scale$($validation$\_$data$[,$ c$("$Income$",$ "Lot Size"$)])$

train$\_$labels $<-$ training$\_$data$Ownership

validation$\_$labels $<-$ validation$\_$data$Ownership

print$($head$($train$\_$features$))$

print$($table$($train$\_$labels$))$

k $<-8$

predictions $<-$ knn$($train $=$ train$\_$features,

test $=$ validation$\_$features,

cl $=$ train$\_$labels,

k $=$ k$)$

accuracy $<-$ mean$($predictions $==$ validation$\_$labels$)*100$

print$($paste$("$Accuracy:$",$ round$($accuracy$,2),"\%"))$

results $<-$ data.frame$($

Actual $=$ validation$\_$labels,

Predicted $=$ predictions,

Income $=$ validation$\_$data$Income,

Lot Size $=$ validation$\_$data$Lot Size

$)$

print$("$KNN Classification Results:"$)$

print$($results$)$

print$($paste$("$Accuracy:$",$ round$($accuracy$,2),"\%"))$

plot$($data$Income, data$Lot$\_$Size,

col $=$ ifelse$($data$Ownership $==$ "owner", "blue", "red"$),$

pch $=$ ifelse$($data$Validation $==0,16,17),$

xlab $=$ "Income $($thousands$)",$

ylab $=$ "Lot Size $($thousands sq$.$ ft$)",$

main $=$ "RidingMowers KNN Classification"$)$

legend$("$topleft$",$

legend $=$ c$("$Owner$",$ "Non$-$owner", "Training", "Validation"$),$

col $=$ c$("$blue$",$ "red", "black", "black"$),$

pch $=$ c$(16,16,16,17))$

validation$\_$labels$\_$numeric$<-$ as$.$numeric$($validation$\_$labels$==$"owner"$)$

predictions$\_$numeric $<-$ as$.$numeric$($predictions$==$"owner"$)$

roc$\_$curve$\_$knn $<-$ roc$($validation$\_$labels, as$.$numeric$($predictions$))$

plot$($roc$\_$curve$\_$knn$,$ main $=$ "ROC Curve for k$-$NN$",$ col $=$ "blue"$)$

pred$\_$knn $<-$ prediction$($as$.$numeric$($predictions$),$

as$.$numeric$($validation$\_$labels$\_$numeric$))$

perf$\_$knn $<-$ performance$($pred$\_$knn$,$ "lift", $"$rpp$")$

plot$($perf$\_$knn$,$ main $=$ "Lift Curve for k$-$NN$",$ col $=$ "red"$)$

actual $<-$ as$.$numeric$($validation$\_$labels $==$ "owner"$)$

predicted$\_$numeric$\_$knn $<-$ as$.$numeric$($predictions $==$ "owner"$)$

rase$\_$knn $<-$ sqrt$($mean$(($actual $-$ predicted$\_$numeric$\_$knn$)^2))$

rsq$\_$knn $<-1-($sum$(($actual $-$ predicted$\_$numeric$\_$knn$)^2)/$ sum$(($actual $-$ mean$($actual$))^2))$

print$($paste$("$RASE for k$-$NN:$",$ rase$\_$knn$))$

print$($paste$("$R$-$squared for k$-$NN:$",$ rsq$\_$knn$))$

predict$\_$ownership $<-$ function$($income$,$lot size$)\{$

new$\_$data $<-$ scale$($matrix$($c$($income$,$ lot size$),$ ncol $=2),$

center $=$ attr$($train$\_$features, "scaled:center"$),$

scale $=$ attr$($train$\_$features, "scaled:scale"$))$

prediction $<-$ knn$($train $=$ train$\_$features,

test $=$ new$\_$data,

cl $=$ train$\_$labels,

k $=$ k$)$

return$($prediction$)$

$\}$

new$\_$customer$\_$prediction $<-$ predict$\_$ownership$($income $=70,$ lot size $=20)$

print$($paste$("$Predicted ownership status:", new$\_$customer$\_$prediction$))$

code for logistic regression

library$($caret$)$

library$($pROC$)$

library$($ROCR$)$

library$($readxl$)$

data $<-$ read$\_$excel$("$C:$/$Users$/$hemun$/$Downloads$/$RidingMowerskNN$.$xlsx$")$

data$Ownership $<-$ factor$($data$Ownership, levels $=$ c$("$non$-$owner", "owner"$))$

set.seed$(123)$

split$\_$index $<-$ createDataPartition$($data$Ownership, p $=0\mathrm{.}7,$ list $=$ FALSE$)$

train $<-$ data$[$split$\_$index, $]$

test $<-$ data$[-$split$\_$index, $]$

log$\_$model $<-$ glm$($Ownership ~ Income $+$ Lot Size, data $=$ train, family $=$ "binomial"$)$

log$\_$pred $<-$ predict$($log$\_$model, test, type $=$ "response"$)$

log$\_$pred$\_$class $<-$ factor$($ifelse$($log$\_$pred $>0\mathrm{.}5,$ "owner", "non$-$owner"$),$

levels $=$ c$("$non$-$owner", "owner"$))$

conf$\_$mat$\_$log $<-$ confusionMatrix$($log$\_$pred$\_$class, test$Ownership$)$

print$($conf$\_$mat$\_$log$)$

roc$\_$curve$\_$log $<-$ roc$($test$Ownership, as$.$numeric$($log$\_$pred$))$

plot$($roc$\_$curve$\_$log$,$ main $=$ "ROC Curve for Logistic Regression", col $=$ "blue"$)$

pred$\_$log $<-$ prediction$($log$\_$pred, test$Ownership$)$

perf$\_$log $<-$ performance$($pred$\_$log$,$ "lift", $"$rpp$")$

plot$($perf$\_$log$,$ main $=$ "Lift Curve for Logistic Regression", col $=$ "red"$)$

actual $<-$ as$.$numeric$($test$Ownership $==$ "owner"$)$

predicted $<-$ as$.$numeric$($log$\_$pred$\_$class $==$ "owner"$)$

rase$\_$log $<-$ sqrt$($mean$(($actual $-$ predicted$)^2))$

rsq$\_$log $<-1-($sum$(($actual $-$ predicted$)^2)/$ sum$(($actual $-$ mean$($actual$))^2))$

print$($paste$("$RASE for Logistic Regression:", rase$\_$log$))$

print$($paste$("$R$-$squared for Logistic Regression:", rsq$\_$log$))$

Give me a code for ensembles with k$-$NN and logistic regression