基于决策树ID3算法的改进

Improved ID3 Algorithm Based on Decision Tree

要针对决策树ID3算法复杂的对数运算以及属性取值多向依赖的缺陷，提出了一种改进的CEID3算法。该方法引入粗糙集论中属性重要度和关联度的概念，并依据这两个概念对决策树ID3算法进行有效改进。仿真结果表明，新算法简化了运算使得终形成的决策树更加符合实际需求。

The complex logarithm operation and multi dependency attribute value is a defect for Decision tree ID3 algorithm, this paper presents an improved algorithm CEID3. The method introduced relation Degree in the Rough Sets, and improve Decision Tree ID3 Algorithm effecti Importance Degree and Cor- vely based on these two con-cepts. The simulation results show this new algorithm simplifies the operation which makes the final decision tree formed more in line with actual needs.