决策树算法在纳税信用等级评定中的应用研究.pdfVIP

ISSN 1009-3044 E-mail: jslt@ 与技术 第 卷第 期 年 月 Computer Knowledge and Technology 电脑知识 5 2 (2009 1 ) Computer Knowledge And Technology 电脑知识与技术 Vol.5,No.2,January 2009, pp.286-288 Tel:+86-551-5690963 5690964 决策树算法在纳税信用等级评定中的应用研究 徐邵兵 （安徽省地方税务局信息中心，安徽合肥 230061 ） 摘要：纳税信用等级评定的实现是需要对大量税收数据进行分析和判定的结果，决策树是进行数据挖掘和分类的常用工具，其中以 C4.5 算法最为流行。 如何应用数据挖掘技术改变纳税信用等级手工评定的现状是当前税务系统税收信息化工作难点之一。 文章主 要讨论如何应用C4.5 算法构造纳税信用等级评定决策树，通过对纳税人涉税数据的采集、预处理、属性选择、决策树生成和剪枝等 一系列过程最终生成纳税信用等级评定决策树，并根据生成的决策树实现对纳税人纳税信用等级的判决。 关键词：决策树；C4.5 算法；纳税信用等级评定 中图分类号： ； 文献标识码： 文章编号： TP182 F810.423 A 1009-3044(2009)02-0286-03 Application Research of Decision Tree Algorithm on Evaluation Ratepaying Credit Grade XU Shao-bing (Information Center of Anhui Province Local Taxation Bureau, Hefei 230061, China) Abstract: The implement of evaluation ratepaying credit grade needs the result of analyzing and judging lots of taxation data, and decision tree is the one of the common tools for data mining and classification, especially C4.5,is a kind of classification algorithm. How to apply da- ta mining technology to change the phenomenon, that evaluation ratepaying credit grade by manual operation, is one of the difficult points of todays tax system taxation informatization work. The main topic of this treatise is discussing how to build a ratepaying credit grade deci- sion tree with C4.5 algorithm, finally build it through a serious of processes, such as gathering, preprocessing, attribute choosing, decision tree generating and pruning, and judge the taxation credit grade by the generated decision tree. Key words: decision tree; C4.5 algorithm; evaluation ratepayi