Cost-Sensitive C4.5 with Post-Pruning and Competition

By Zilong Xu et al

Published on Aug. 21, 2018

Read the original document by opening this link in a new tab.

Table of Contents

1. Introduction
2. Preliminaries
3. The Algorithm
3.1. The Heuristic Information
3.2. Post-Pruning Technique
3.3. The Competition Approach
4. A Running Example

Summary

Decision tree is an effective classification approach in data mining and machine learning. This paper presents a cost-sensitive decision tree algorithm inspired by C4.5 for numeric data. The algorithm focuses on test cost weighted information gain ratio and post-pruning strategy to reduce total cost. Experimental results show the effectiveness of the algorithm in generating stable and cost-sensitive decision trees.