基于CART决策树数据挖掘算法的应用研究.pdf

第30卷第10 期 煤 炭 技 术 Vol.30,No.10 2011 年10 期 Coal Technology October,2011 基于CART 决策树数据挖掘算法的应用研究 1 2 陈辉林 ，夏道勋 （1. 贵州师范大学物理与电子科学学院，贵阳 550001；2. 贵州师范大学贵州省信息与计算科学重点实验室，贵阳 550001） 摘 要:分类与回归树CART 算法是数据挖掘技术中重要的算法。依据CART 算法理论，采用类型变量求解决策 树，并引入优化的分裂函数，然后利用基于类型变量的论域划分创建二叉树，抽取和筛选预测准则，从而为职能部 门决策提供科学而可靠的依据。最后以贵州师范大学教学与管理中的数据，给出算法的应用实例。 关键词:CART；决策树；类型变量；数据挖掘 中图分类号:TP301.6 文献标识码:A 文章编号:1008-8725 （2011）10-0164-03 Applied Research on Data Mining Based on CART Decision Tree Algorithm 1 2 CHEN Hui-lin , XIA Dao-xun （1. School of Computer Science and Technology Guizhou University, Guiyang 550001, China; 2. Key Laboratory of Information and Computing Science of Guizhou Province, Guizhou Normal University, Guiyang 550001, China） Abstract:CART－Classification and Regression Trees algorithm is an important algorithm in Data Mining. In this paper, we based on the theory of CART algorithm, adopted the algorithm of solving the decision tree under the categorical variables, introduced the optimized splitting function, and then made use of the division of the universe of discourse in type variable to construct binary tree, extracted and screened the forecasting criterion. The purpose of this paper is to provide a scientific and reliable basis for the department decision. At last, the application example of the algorithm was provided with the data in teaching and administration of GuiZhou normal university in this paper. Key words:CART; decision tree; type variable; data mining 算法、Naive