Classification Using Decision Trees

Abstract

This chapter presents the most widespread ensemble method, the decision tree. A decision tree classifier estimates a categorical dependent variable or a continuous dependent. It solves binary and multiclass classification problems. We base the model on a tree-like structure. It breaks down the data into small, manageable chunks while incrementally developing a decision tree. The outcome is a tree-like structure with decision nodes and leaf nodes. We consider it a greedy model since its primary concern is to reduce the training time while maximizing information gain.