一种基于属性贡献度的决策树学习算法  

An algorithm of decision tree learning based on attribute contribution

作　　者：孙淮宁[1,2] 胡学钢[1] 

机构地区：[1]合肥工业大学计算机与信息学院,安徽合肥230009 [2]淮南师范学院计算机与信息工程系,安徽淮南232001

出　　处：《合肥工业大学学报（自然科学版）》2009年第8期1137-1141,共5页Journal of Hefei University of Technology：Natural Science

基　　金：安徽省高等学校自然科学研究基金资助项目(2006KJ164C)

摘　　要：文章提出一种基于属性贡献度的决策树算法ICD,该算法利用信息熵理论构建基于样本类别信息分布确定的属性贡献度函数模型,从属性选优上消除了多值属性选择的偏向,将ICD算法与ID3、C4.5算法进行实验对比,数据结果表明该算法具有良好的性能。An algorithm of decision tree learning based on attribute contribution, which is called the ICE） algorithm, is proposed. The algorithm constructs the model of attribute contribution functions based on certain classic information distributing by the information entropy theory. It eliminates the bias of selecting the attribute with many values by the best attribute choice. The comparison among the algorithms ICD, ID3 and C4. 5 is made by experiments, and the results indicate that the ICD algorithm has better performance.

关 键 词：决策树 归纳学习 属性贡献度 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]