Analysis of CART Algorithms in Data Mining
Junqi Sun, Nianzu Jiang, Guang Sun, Wenbing Huang · 2023
In the information age, with the continuous growth of massive data, the demand for processing data to obtain information in various fields is increasing. Among various data mining methods, CART, the classification and regression tree, stands out as an intuitive and efficient decision tree algorithm. It is composed of feature selection, tree generation, and pruning. Gini is selected as the index in classification and MSE in regression. After recursively constructing a binary tree, use the criteria to pruning, and finally, a good tree is obtained. Moreover, in section 3, we exemplify CART is used for selecting the best features to classify data, whether it is discrete data or continuous data. In addition, we use UCI data to conduct experiments to demonstrate that the CART algorithm after feature selection has a higher dataset classification accuracy rate. From this paper, what can be found is that the CART algorithm is useful, and it is worth combining other methods to continue to study more applications.