改进的基于排序熵的有序决策树算法-模式识别与人工智能.PDFVIP

第27 卷摇 第2 期摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 模式识别与人工智能摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 Vol.27摇 No.2 摇 2014年2月摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 PR AI摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 Feb.摇 2014 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 改进的基于排序熵的有序决策树算法* 陈建凯摇 摇 王熙照摇 摇 高相辉 (河北大学 数学与计算机学院河北省机器学习与计算智能重点实验室摇 保定071002) 摘摇 要摇 由于基于排序熵的有序决策树在扩展属性选取时,需计算每个条件属性的每个割点处的排序互信息,并 通过对比这些排序互信息的大小来确定最大值(最大值对应的属性为扩展属性),计算复杂度较高.针对此问题,文 中将割点分为平衡割点和非平衡割点两部分,建立一个数学模型,从理论上证明排序互信息最大值不会在平衡割 点处达到,而只能在非平衡割点处达到.这说明在计算排序互信息时只需遍历非平衡割点,而无需再计算平衡割点 处的值,从而使决策树构建的计算效率得到较大程度提高.数值实验验证此结果. 关键词摇 有序分类,有序决策树,非平衡割点 中图法分类号摇 TP 181 Improved Ordinal Decisions Trees Algorithms Based on Rank Entropy CHENJian鄄Kai,WANG Xi鄄Zhao,GAO Xiang鄄Hui (Key Laboratory onMachine Learning and Computational Intelligence of HebeiProvince, College of Mathematics and Computer Science,Hebei University,Baoding 071002) ABSTRACT When the expanded attributes are selected for decision tree learning based on rank entropy, computing the rank mutual information of every single cut for each of the continuous鄄valued attributes is required togettheexpandedattributeby comparingthevaluesofrankmutualinformation. Therefore,the computational complexity ishigh. Aiming atthisproblem,cut鄄pointsaredividedinto stableandunstable cut鄄pointsandamathematicalmodelisestablishedinthispaper. Theproposedmodeltheoreticallyproves that the rank mutual information function achieves its maximum not at stable cut鄄points,but at unstable cut鄄points. The result means that in the algorithm only traversing the unstable cut鄄points is required instead of computing the values of the stable cut鄄points. Thus,the computational efficiency of