Classification and Regression Tree CART Decision Tree
CART Decision Tree is a decision tree algorithm that can be used for classification and regression tasks. It was developed in 1977 by Leo Breiman and Charles Joel Stone, professors of statistics at Berkeley University, and Jerome H. Friedman and Richard Olshen at Stanford University. The algorithm is widely used in the fields of data mining and machine learning. Its core is to recursively split the data set into smaller and smaller subsets and build a decision tree in the process.
CART generates a binary tree, each internal node has a decision rule, and each leaf node gives a prediction result. Each split is based on the threshold of a single variable. CART uses pruning strategies to avoid overfitting and improve the generalization ability of the model. CART can also build classification trees and regression trees, which are suitable for different types of prediction tasks.