# Decision Tree Classification

# we may need to prune the DT to avoid overfitting

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))

# 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
# DT does not require scaling because it does not use Euclidean distance.
# However, if we dont scale the data, the visualization will take much longer.
# I must mention that I noticed the performance of DT with scaled data is degraded slightly.
training_set[-3] = scale(training_set[-3])
test_set[-3] = scale(test_set[-3])

# Fitting Decision Tree Classification to the Training set
# install.packages('rpart')
library(rpart)
# DT without pruning
classifier <- rpart(formula = Purchased ~ .,
                   data = training_set,
                   method = "class",
                   control = rpart.control(minsplit = 1))

# DT with post-pruning
pruned_DT <- prune(classifier,
                   cp = classifier$cptable[which.min(classifier$cptable[,"xerror"]),"CP"])

# Predicting the Test set results
# The below code will give the probabilities
# y_pred = predict(classifier, newdata = test_set[-3])

# The below code will give the class
y_pred = predict(classifier, newdata = test_set[-3], type = 'class')

# Making the Confusion Matrix
y_pred <- factor(y_pred, levels = c(0,1), labels = c(0,1)) # using class lables
cm <- as.matrix(table(Actual = test_set[, 3], Predicted = y_pred)) # create the confusion matrix

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

# Model Evaluation
# Applying k-Fold Cross Validation
# The DT without prunning
# install.packages('caret')
library(caret)
folds = createFolds(training_set$Purchased, k = 10)
cv = lapply(folds, function(x) {
  training_fold = training_set[-x, ]
  test_fold = training_set[x, ]
  classifier <- rpart(formula = Purchased ~ .,
                      data = training_fold,
                      method = "class",
                      control = rpart.control(minsplit = 1))
  y_pred = predict(classifier, newdata = test_fold[-3], type = 'class')
  y_pred <- factor(y_pred, levels = c(0,1), labels = c(0,1))
  cm = table(test_fold[, 3], y_pred)
  #accuracy = (cm[1,1] + cm[2,2]) / (cm[1,1] + cm[2,2] + cm[1,2] + cm[2,1])
  accuracy = sum(diag(cm)) / sum(cm)
  return(accuracy)
})
(accuracy_k_folds <- mean(as.numeric(cv)))

# Model Evaluation
# Applying k-Fold Cross Validation
# The DT after prunning
library(caret)
folds = createFolds(training_set$Purchased, k = 10)
cv = lapply(folds, function(x) {
  training_fold = training_set[-x, ]
  test_fold = training_set[x, ]
  pruned_DT <- prune(classifier,
                     cp = classifier$cptable[which.min(classifier$cptable[,"xerror"]),"CP"])
  y_pred = predict(pruned_DT, newdata = test_fold[-3], type = 'class')
  y_pred <- factor(y_pred, levels = c(0,1), labels = c(0,1))
  cm = table(test_fold[, 3], y_pred)
  #accuracy = (cm[1,1] + cm[2,2]) / (cm[1,1] + cm[2,2] + cm[1,2] + cm[2,1])
  accuracy = sum(diag(cm)) / sum(cm)
  return(accuracy)
})
(accuracy_k_folds <- mean(as.numeric(cv)))


# Applying Grid Search to find the best parameters (error loading package!)
# Logistic Model Trees
# install.packages('caret')
library(caret)
classifier = train(form = Purchased ~ ., data = training_set, method = 'LMT')
classifier
classifier$bestTune


# Visualising the Training set results
library(ElemStatLearn)
set = training_set
X1 = seq(min(set[, 1]) - 1, max(set[, 1]) + 1, by = 0.01)
X2 = seq(min(set[, 2]) - 1, max(set[, 2]) + 1, by = 0.01)
grid_set = expand.grid(X1, X2)
colnames(grid_set) = c('Age', 'EstimatedSalary')
y_grid = predict(classifier, newdata = grid_set, type = 'class')
plot(set[, -3],
     main = 'Decision Tree Classification (Training set)',
     xlab = 'Age', ylab = 'Estimated Salary',
     xlim = range(X1), ylim = range(X2))
contour(X1, X2, matrix(as.numeric(y_grid), length(X1), length(X2)), add = TRUE)
points(grid_set, pch = '.', col = ifelse(y_grid == 1, 'springgreen3', 'tomato'))
points(set, pch = 21, bg = ifelse(set[, 3] == 1, 'green4', 'red3'))

# Visualising the Test set results
library(ElemStatLearn)
set = test_set
X1 = seq(min(set[, 1]) - 1, max(set[, 1]) + 1, by = 0.01)
X2 = seq(min(set[, 2]) - 1, max(set[, 2]) + 1, by = 0.01)
grid_set = expand.grid(X1, X2)
colnames(grid_set) = c('Age', 'EstimatedSalary')
y_grid = predict(classifier, newdata = grid_set, type = 'class')
plot(set[, -3], main = 'Decision Tree Classification (Test set)',
     xlab = 'Age', ylab = 'Estimated Salary',
     xlim = range(X1), ylim = range(X2))
contour(X1, X2, matrix(as.numeric(y_grid), length(X1), length(X2)), add = TRUE)
points(grid_set, pch = '.', col = ifelse(y_grid == 1, 'springgreen3', 'tomato'))
points(set, pch = 21, bg = ifelse(set[, 3] == 1, 'green4', 'red3'))

# Plotting the tree
# These generate poor plots
plot(classifier)
text(classifier)

# A better approach to plot DT

library(rpart)
library(rpart.plot)

prp(classifier)
prp(pruned_DT)

rpart.plot(classifier)
rpart.plot(pruned_DT)

# Diplaying and ploting Complexity Parameter (cp)
printcp(classifier)
plotcp(classifier)
rsq.rpart(classifier)