Changing the prior probabilities

As mentioned in the video, you can also change the prior probabilities to obtain a decision tree. This is an indirect way of adjusting the importance of misclassifications for each class. You can specify another argument inside rpart() to include prior probabities. The argument you are looking for has the following form

parms = list(prior=c(non_default_proportion, default_proportion))

The rpart package is now already loaded in your workspace.

This exercise is part of the course

Credit Risk Modeling in R

View Course

Exercise instructions

Change the code provided such that a decision tree is constructed , including the argument parms and changing the proportion of non-defaults to 0.7, and of defaults to 0.3 (they should always sum up to 1). Additionally, include control = rpart.control(cp = 0.001) as well.
Plot the decision tree using the function plot and the tree object name. Add a second argument "uniform=TRUE" to get equal-sized branches.
Add labels to the tree using function text() and the decision tree object name.

Hands-on interactive exercise

Have a go at this exercise by completing this sample code.

# Change the code below such that a tree is constructed with adjusted prior probabilities.
tree_prior <- rpart(loan_status ~ ., method = "class",
                    data = training_set)

# Plot the decision tree


# Add labels to the decision tree

Edit and Run Code