Improvement and Application of Decision Tree C4.5 Algorithm

Abstract

The C4.5 algorithm as the most popular decision tree algorithm, there are still some deficiencies, C4.5 algorithm uses post-pruning to solve the over-fitting problem, but increase the modeling overhead, in response to this problem the idea of combining over-fitting branches ahead of time in the process of creating a decision tree is proposed and improved on the original C4.5 algorithm, which solves the problem of long decision-making time and large models of the decision tree. Experimental verification shows that the improved algorithm has a significant improvement in model simplification and model accuracy. The improved algorithm is applied to the "Engineering Quality Decision Support System" to provide supervision and decision-making support for the supervisory department.