一种基于变精度粗糙集的C4.5决策树改进算法  

Improved C4.5 decision trees algorithm based on variable precision rough set

作　　者：刘兴文[1] 王典洪[1] 陈分雄[1] 

机构地区：[1]中国地质大学,武汉430074

出　　处：《计算机应用研究》2011年第10期3649-3651,共3页Application Research of Computers

基　　金：湖北省自然科学基金资助项目(2010CDB04203);国家教育部长江三峡库区地质灾害研究中心开放基金资助课题(TGRC201018)

摘　　要：针对C4.5决策树构造复杂、分类精度不高等问题,提出了一种基于变精度粗糙集的决策树构造改进算法。该算法采用近似分类质量作为节点选择属性的启发函数,与信息增益率相比,该标准更能准确地刻画属性分类的综合贡献能力,同时对噪声有一定的抑制能力。此外还针对两个或两个以上属性的近似分类质量相等的特殊情形,给出了如何选择最优的分类属性作为节点的方法。实验结果证明,该算法构造的决策树在分类精度和规模上均优于C4.5算法。Aiming at the problems of complexisity and relatively low classification accuracy of decision trees constructed by C4.5 algorithm,this paper proposed a new decision trees classification algorithm（VPRSC4.5） based on the variable precision rough set（VPRS）,which took the approximate quality of classification as the heuristic function in order to alleviate the effect of noise data on choosing splitting attributes.It also gave out the solution to the problem how to choose the best attributes as the node when two or more attributes had the same value of approximate quality of classification.Experiments prove that the size and classification accuracy of the decision trees generated by the improved algorithm is superior to the C4.5 algorithm.

关 键 词：数据挖掘 决策树 信息增益率 C4.5算法 粗糙集 变精度粗糙集 近似分类质量 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]