决策树算法的比较与应用研究  

Comparison and Application of Decision Tree Algorithm

作　　者：张海燕[1] 刘岩[2] 马丽萌 苑津莎[1] 巨汉基[2] 魏彤珈 Zhang Haiyan Liu Yan Ma Limeng Yuan Jinsha Ju Hanji Wei Tongjia(North China Eleetrie Power University, Baoding 071003, China State Grid Jibei Eleetrie Power Co. Ltd. Researeh Institute, North China Electric Power Research Institute Co. Ltd., Beijing 100045 ,China)

机构地区：[1]华北电力大学,河北保定071003 [2]国网冀北电力有限公司电力科学研究院(华北电力科学研究院有限责任公司),北京100045

出　　处：《华北电力技术》2017年第6期42-47,共6页North China Electric Power

摘　　要：数据挖掘中一项重要的任务是对数据进行分类,而决策树是分类算法中一个最主要而且应用最广泛的算法分支。文章叙述了决策树的工作原理和基本概况,介绍了几种常用的决策树算法,并通过对5种公开数据集的分类结果进行比较,验证了这些算法的优劣,最后将随机森林算法应用到电力用户信用评价中,为进一步研究提供了思路。An important task in data mining is to classify the data, and the decision tree is one of the most important and widely used algorithm branches in the classification algorithm. This paper describes the working principle and basic situation of the decision tree, and introduces several commonly used decision tree algorithms, and compares the classification results of five public data sets to verify the advantages and disadvantages of these algorithms. Finally the Random Forest algorithm is applied to the credit evaluation of the electricity user ,which provides the idea for further research.

关 键 词：数据挖掘 分类算法 决策树 

分 类 号：TM769[电气工程—电力系统及自动化]