# ROC
# using Logistic Regression as an example

rm(list = ls())

# Importing the dataset
dataset <- read.csv(file.path(getwd(),'Data/Social_Network_Ads.csv'))
dataset <- dataset[3:5]

# Encoding the target feature as factor
dataset$Purchased <- factor(dataset$Purchased, levels = c(0, 1), labels = c(0, 1))

# Splitting the dataset into the Training set and Test set
# install.packages('caTools')
library(caTools)
set.seed(123)
split <- sample.split(dataset$Purchased, SplitRatio = 0.75)
training_set <- subset(dataset, split == TRUE)
test_set <- subset(dataset, split == FALSE)

# Feature Scaling
training_set[-3] <- scale(training_set[-3])
test_set[-3] <- scale(test_set[-3])

# Fitting Logistic Regression to the Training set
classifier <- glm(formula = Purchased ~ .,
                  family = binomial,
                  data = training_set)

# Predicting the Test set results
prob_pred <- predict(classifier, type = 'response', newdata = test_set[-3])
y_pred <- ifelse(prob_pred > 0.5, 1, 0)
y_pred <- factor(y_pred, levels = c(0,1), labels = c(0,1))

# Making the Confusion Matrix
# cm = table(test_set[, 3], y_pred)
# cm = table(test_set[, 3]  > 0.5 , y_pred > 0.5) we cant use > when having factors
cm <- as.matrix(table(Actual = test_set[, 3], Predicted = y_pred)) # create the confusion matrix

(accuracy <- mean(y_pred == test_set$Purchased))

# Model Evaluation
# Using Receiver Operating Characterstic (ROC) Curve
# install.packages("ROCR")
library(ROCR)
roc_pred <- prediction(prob_pred, test_set[3])
roc_eval <- performance(roc_pred, "tpr", "fpr")
plot(roc_eval,
     colorize = TRUE)

# Add another curve
# plot(perf2, add = TRUE, colorize = TRUE)
# then, instead of using colorize = TRUE, we can use col = as.list(1:10)

# Add lables to each curve
text(locator(), labels = c("Logistic Regression"), cex = 0.7) # clicking on an existing graph, then click on finish.
# coor <- (locator()) to get the coordiante of x and y axes
# text(x = 0.5, y = 0.5, labels = c("Logistic Regression"))
# or
# text(coor$x, coor$y, labels = c("Logistic Regression"))

abline(a = 0, b = 1)

# Area Under the Curve (AUC)
lr_auc <- performance(roc_pred, "auc")
lr_auc <- unlist(slot(auc, "y.values"))
lr_auc <- round(auc, 3)
lr_auc <- paste(c("Logistic Regression "),lr_auc, sep="")
# dt_auc <- paste(c("Decision Tree "), dt_auc, sep="")

# legend(.6, .4, auc, title = "AUC", cex = 0.5)
# legend("bottomright", paste(auc), col = c("green"), pch = c(3), cex = 0.5)
# legend(0.6, 0.2, title = "AUC", c(lr_auc), col = c("green"), pch = c(3), cex = 0.5, box.col = "white", border="white")
legend(0.6, 0.2, title = "AUC", c(lr_auc), col = c("green"), cex = 0.6, box.col = "white",
       border="white")