Exercises

1. What is the main difference between random forests and gradient-boosted trees?
2. Explain why the Gini Impurity may be interpreted as the misclassification rate.
3. Explain why it is necessary to perform feature encoding for categorical features.
4. In this chapter, we provided two ways to do feature encoding. Find one other way to encode categorical features.
5. Explain why the accuracy metric we used in Chapter 2, Classifying Twitter Feeds with Naive Bayes, is not suitable for predicting clicks on our dataset.
6. Find other objectives we can use for the XGBoost algorithm. When would you use each objective?