Data Classification using the Random Forest Method

Abstract

This paper focuses on the problem of data clas- sification using decision trees and Random Forests. The theoretical background of de- cision trees, entropy and information gain as the basic principles of classification is described. Sub- sequently, a custom decision tree-based classifier is implemented as well as an ensemble Random For- est model that improves robustness and accuracy through bagging. The task of the models was to predict the ticket price category based on various attributes. The results are evaluated using accu- racy and confusion matrices, with Random Forest demonstrating higher classification accuracy com- pared to the standalone decision tree.