IMPORTING DATA

Reading Data

library(ISLR)
# reading inbuilt data as data frame
default.df <- as.data.frame(Default)
# attach data frame
attach(default.df)
# dimension of the data frame
dim(default.df)

[1] 10000     4

Data Structure

# structure of the data table
str(default.df)

'data.frame':   10000 obs. of  4 variables:
 $ default: Factor w/ 2 levels "No","Yes": 1 1 1 1 1 1 1 1 1 1 ...
 $ student: Factor w/ 2 levels "No","Yes": 1 2 1 1 1 2 1 2 1 1 ...
 $ balance: num  730 817 1074 529 786 ...
 $ income : num  44362 12106 31767 35704 38463 ...

Descriptive Statistics

# descriptive statistics of the dataframe
library(psych)
describe(default.df)[, c(1:5)]

         vars     n     mean       sd   median
default*    1 10000     1.03     0.18     1.00
student*    2 10000     1.29     0.46     1.00
balance     3 10000   835.37   483.71   823.64
income      4 10000 33516.98 13336.64 34552.64

SPLITTING DATA

(DATA TRAINING AND TESTING)

BUILDING MODELS

(LOGIT AND DECISION TREE)

Control Parameters

# control parameters
objControl <- trainControl(method = "boot", 
                           number = 2, 
                           returnResamp = 'none', 
                           summaryFunction = twoClassSummary, 
                           classProbs = TRUE,
                           savePredictions = TRUE)

Building Logit Model using caret Package

# model building using caret package
set.seed(766)
caretLogitModel <- train(trainData.df[, 2:4],
                      trainData.df[, 1],
                      method = 'glmStepAIC',
                      trControl = objControl,
                      metric = "ROC",
                      verbose = FALSE)

Building Tree Model using caret Package (Information Gain)

# decision tree using method criterian information gain
dTreeInfoGain <- train(default ~ ., 
                       data = trainData.df, 
                       method = "rpart", 
                       parms = list(split = "information"), 
                       trControl = trainControl(method = "cv"))
dTreeInfoGain

Building Random Forest Tree Model using caret Package

# setting parameters
library(caret)
ctrl <- trainControl(method = "repeatedcv", 
                     number = 2, 
                     repeats = 2, 
                     selectionFunction = "oneSE")

set.seed(766)
# fitting random forest model
modelRF <- train(default ~ ., 
                 data = testData.df, 
                 method = "rf", 
                 metric = "Kappa",
                 trControl = ctrl)
modelRF

PREDICTED PROBABILITIES PLOTS

(LOGIT, DECISION TREE AND RANDOM FOREST)

Predicted Probabilities of Default (Yes / No), based on Test Data (Logit)

# predicted probabilities
predTestProb <- predict(caretLogitModel, testData.df, type = "prob")
# plot of probabilities
plot(predTestProb[,2], 
     main = "Scatterplot of Probabilities of Default (test data)", 
     xlab = "Customer ID", ylab = "Predicted Probability of Default")

Predicted Probabilities of Default (Yes / No), based on Test Data (Tree)

# predicted probabilities
predTestProbInfoGain <- predict(dTreeInfoGain, testData.df, type = "prob")
# plot of probabilities
plot(predTestProbInfoGain[,2], 
     main = "Scatterplot of Probabilities of Default (test data)", 
     xlab = "Customer ID", ylab = "Predicted Probability of Default")

Predicted Probabilities of Default (Yes / No), based on Test Data (Random Forest)

# predicted probabilities
predTestProbRF <- predict(modelRF, testData.df, type = "prob")
# plot of probabilities
plot(predTestProbRF[,2], 
     main = "Scatterplot of Probabilities of Default (test data)", 
     xlab = "Customer ID", ylab = "Predicted Probability of Default")

PREDICTION AND CONFUSION MATRIX

(LOGIT, DECISION TREE AND RANDOM FOREST)

Confusion Matrix based on Test Data (Logit)

# confusion matrix (logit)
confusionMatrix(predClass, testData.df$default, positive = "Yes")

Confusion Matrix based on Test Data (Tree)

# confusion matrix (decision tree)
confusionMatrix(predClassInfoGain, testData.df$default, positive = "Yes")

Confusion Matrix based on Test Data (Random Forest)

# confusion matrix (random forest)
confusionMatrix(predClassRF, testData.df$default, positive = "Yes")

ROC CURVES AND AUC

(LOGIT, DECISION TREE AND RANDOM FOREST)

ROC Plot on the Test data (Logit)

library(ROCR)
# prediction
PredictObjectLogit <- prediction(predTestProb[2], testData.df$default)

# performance
PerformObjectLogit <- performance(PredictObjectLogit, "tpr","fpr")

# Plot the ROC Curve for Credit Card Default
plot(PerformObjectLogit, 
     main = "ROC Curve for CC Default (Logit)",
     col = "red",
     lwd = 2)
abline(a = 0,b = 1,lwd = 2,lty = 3,col = "black")

ROC Plot on the Test data (Tree)

library(ROCR)
# prediction
PredictObjectTree <- prediction(predTestProbInfoGain[2], testData.df$default)

# performance
PerformObjectTree <- performance(PredictObjectTree, "tpr","fpr")

# Plot the ROC Curve for Credit Card Default
plot(PerformObjectTree, 
     main = "ROC Curve for CC Default (Decision Tree)",
     col = "blue",
     lwd = 2)
abline(a = 0,b = 1,lwd = 2,lty = 3,col = "black")

ROC Plot on the Test data (Random Forest)

library(ROCR)
#Every classifier evaluation using ROCR starts with creating a prediction object. This function is used to transform the input data into a standardized format.
PredictObjectRF <- prediction(predTestProbRF[2], testData.df$default)

# All kinds of predictor evaluations are performed using the performance function
PerformObjectRF <- performance(PredictObjectRF, "tpr","fpr")

# Plot the ROC Curve for Credit Card Default
plot(PerformObjectRF, 
     main = "ROC Curve for CC Default (Random Forest)",
     col = "green",
     lwd = 2)
abline(a = 0,b = 1,lwd = 2,lty = 3,col = "black")

Comparing ROC Curves of Classifiers (Logit, Tree and Random Forest)

# ROC curves (logit, tree and random forest)
plot(PerformObjectLogit, col = "red", lwd = 2,
     main = "ROC Curve for CC Default \n (Logit, Decision Tree and Random Forest)")
plot(PerformObjectTree, add = TRUE, col = "blue", lwd = 3,
     main = "ROC Curve for CC Default \n (Logit, Decision Tree and Random Forest)")
plot(PerformObjectRF, add = TRUE, col = "green", lwd = 3,
     main = "ROC Curve for CC Default \n (Logit, Decision Tree and Random Forest)")