In order to properly evaluate the performance of a classification tree on these data, we must estimate the test error rather than simply computing the training error. We split the observations into a training set and a test set, build the tree using the training set, and evaluate its performance on the test data. The predict() function can be used for this purpose. In the case of a classification tree, the argument type="class" instructs R to return the actual class prediction. This approach leads to correct predictions for 68% of the locations in the test data set.

set.seed(3)
train <- sample(1:nrow(Carseats), 200)
Carseats.test <- Carseats[-train,]
High.test <- High[-train]
tree.carseats <- tree(High ~ . - Sales, Carseats, subset = train)
tree.pred <- predict(tree.carseats, Carseats.test, type = "class")
table(tree.pred, High.test)
         High.test
tree.pred No Yes
      No  88  41
      Yes 23  48
      
(88 + 48) / 200
[1] 0.68

Questions

• Create a 70-30 train-test split of the OJ data set. Use a seed value of 1. Store the numeric vector with training data indices in train.idx.
• Fit a classification tree on the training data with Purchase as dependent variable and all other variables as independent variables. Store the model in the variable tree.oj.
• Use the tree to make class predictions on the test set. Store the predictions in the variable tree.pred.
• Calculate the confusion matrix. Store the result in the variable cf.test.
  (note1: give the predictions as the first argument in the table() function, and the ground truth as the second argument).
  (note2: write the code for the ground truth directly in the table() function, so WITHOUT using some Purchase.test intermediary variable)
• Finally, calculate the test set accuracy. Store the result in the variable acc.test.
  (note: try to avoid hardcoding and use e.g. the sum() and diag() functions on the confusion matrix)

Assume that:

• The ISLR2 and tree libraries have been loaded
• The OJ dataset has been loaded and attached