决策树分类算法C4.5中连续属性过程处理的改进  

Improvement of Continuous Variables Processing with C4.5 Algorithm

作　　者：李慧慧 万武族[2] 

机构地区：[1]贵州人民武装学院信息工程系,贵州贵阳550025 [2]贵州大学计算机科学系,贵州贵阳550025

出　　处：《计算机与现代化》2010年第8期8-10,共3页Computer and Modernization

基　　金：贵州省省长基金资助项目(200404);贵州大学自然科学青年基金资助项目(2009021)

摘　　要：决策树分类算法C4.5是数据挖掘中最常用、最经典的分类算法。但是C4.5算法也存在一些不足之处,针对C4.5算法处理连续属性比较耗时的特点,本文对连续的处理过程进行改进,以提高算法的计算效率。改进的C4.5算法与原C4.5算法相比,在构造决策树时具有相同的准确率和更高的计算速度。The decision tree classification algorithm C4.5 is the most popular and classical classification algorithm in the data mining.But,there are some defects in it,the processing of continuous variables in the C4.5 algorithm consumes too much time,according to this characteristic,the paper improves the processing of continuous variables to enhance the efficiency of the algorithm.The improved algorithm has better efficiency and has the same accuracy comparing with the C4.5 algorithm when building decision tree.

关 键 词：数据挖掘 决策树 C4.5算法 连续属性 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]