Worksheet

Author

Affiliation

Published

January 19, 2026

1 CART on Insurance Data

Exercise 1 The dataset contains 2220 observations of young drivers with car insurance contracts. The variables are TYPE, VALUE, SEX, AGEV, and AGEI. Below is a step-by-step solution in R.

Load the dataset and describe it.
Split the dataset into training and test samples (80%-20%).
Use the CART algorithm in order to explain the variable TYPE by the variables VALUE, SEX, AGEV and AGEI.

Remark

Remark 1. Ensure TYPE is a factor for classification trees.

Prune the tree using cross-validation (x-error) and the 1-SE rule.
Predict the TYPE on test set and assess prediction quality.
Compute and visualize ROC curves.
Interpret the results and compare them with Linear Discriminant Analysis.

---
date: 2026-01-19
format:
  html:
    title: Worksheet
    code-tools:
      source: true
  pdf:
    title: |
      ![](/images/tse-large.png){height=1.5cm} \
      Worksheet 3 -- Classification and Regression Trees (solution)
    include-before-body:
      - text: |
          \counterwithout{exercise}{section}
---

## CART on Insurance Data

::: {#exr-insurance-cart}

The dataset contains **2220 observations** of young drivers with car insurance contracts. The variables are `TYPE`, `VALUE`, `SEX`, `AGEV`, and `AGEI`. Below is a step-by-step solution in **R**.

1. Load the dataset and describe it.

2. Split the dataset into training and test samples (80%-20%).

3. Use the CART algorithm in order to explain the variable `TYPE` by the variables `VALUE`, `SEX`, `AGEV` and `AGEI`.

:::{#rem-type}
Ensure `TYPE` is a factor for classification trees.
:::

4. Prune the tree using cross-validation (x-error) and the 1-SE rule.

5. Predict the `TYPE` on test set and assess prediction quality.

6. Compute and visualize ROC curves.

7. Interpret the results and compare them with Linear Discriminant Analysis.

:::