Week 6 Exercise

Victor Enchautegui

11/1/2020

Load libraries

library(ggplot2)
library(ROCR) 
library(e1071)

Training set

cc <- read.csv("data/UCI_Credit_Card.csv")

Transform data

cc$default.payment.next.month <- factor(cc$default.payment.next.month,levels=c(0,1), labels=c("No","Yes"))

cc$AGE <- cut(cc$AGE,breaks=c(0,20,30,40,50,60,70,80,100), labels=c("<20s","20s", "30s", "40s", "50s", "60s", "70s", "80s+")) 

cc$AGE <- factor(cc$AGE) 

cc$EDUCATION <- factor(cc$EDUCATION,levels=c(1,2,3,4,5,6), labels=c("Grad School", "University", "High School", "Others", "Unknown", "Unknown"))

cc$PAY_0  <- factor(cc$PAY_0 , levels=c(-2,-1,0,1,2,3,4,5,6,7,8,9), labels=c("Pay duly", "1mo", "2mos", "3mos", "4mos",  "5mos", "6mos", "7mos",  "8mos", "9mos", "10mos", "11mos"))

cc$PAY_2  <- factor(cc$PAY_2 , levels=c(-2,-1,0,1,2,3,4,5,6,7,8,9), labels=c("Pay duly", "1mo", "2mos", "3mos", "4mos",  "5mos", "6mos", "7mos",  "8mos", "9mos", "10mos", "11mos"))

cc$PAY_3  <- factor(cc$PAY_3 , levels=c(-2,-1,0,1,2,3,4,5,6,7,8,9), labels=c("Pay duly", "1mo", "2mos", "3mos", "4mos",  "5mos", "6mos", "7mos",  "8mos", "9mos", "10mos", "11mos"))

Training set

train <- cc[sample(nrow(cc), 100), ] 

Test set

test <- cc[sample(nrow(cc), 100), ]

Build the Naïve Bayes classifier

nbPay <- naiveBayes(default.payment.next.month ~ PAY_0 + PAY_2 + PAY_3, train)

Perform on the testing set

nb_prediction <- predict(nbPay,
                         # remove column "default payment next month"
                         test[,-ncol(test)],
                         type='raw')


score <- nb_prediction[, c("Yes")] 

actual_class <- train$default.payment.next.month == 'Yes' 

pred <- prediction(score, actual_class)

perf <- performance(pred, "tpr", "fpr") 

plot(perf, lwd=2, xlab="False Positive Rate (FPR)",
     ylab="True Positive Rate (TPR)")
abline(a=0, b=1, col="gray50", lty=3)

auc <- performance(pred, "auc")
auc <- unlist(slot(auc, "y.values"))
auc

[1] 0.4528898

Discussion

When AUC is approximately 0.5, model has no discrimination capacity to distinguish between positive class and negative class.

So in my training set the AUC is .45, which indicates a mistake in the way I picked my classifier targets or I might have a bad training set. And being my AUC is close to .5, this suggests no discrimination, so we might as well flip a coin.