决策树算法在电路故障诊断中的应用  

Application of decision tree in circuit fault diagnosis

作　　者：陈绍炜[1] 王聪[1] 赵帅[1] 

机构地区：[1]西北工业大学电子信息学院,西安710129

出　　处：《计算机工程与应用》2013年第12期233-236,共4页Computer Engineering and Applications

基　　金：航空科学基金(No.2012ZD53051)

摘　　要：针对PHM(Prognostic and Health Management)中数据挖掘和知识获取困难的问题,提出一种以J48决策树算法为基础的故障诊断方法。采用了开源数据挖掘软件Weka,对CTSV滤波器故障仿真数据进行计算,对故障数据进行属性清理和参数选择。生成的决策树模型有很高的交叉验证率和分类效果。According to the data mining and knowledge acquisition problems of PHM （Prognostic and Health Management）, this paper proposes a fault diagnosis method based on J48 decision tree algorithm. It calculates the fault simulation data of CTSV filter by using Weka which is a kind of open-source data mining software, then deans up attributes and selects parameters of the aircraft engine. The decision tree model has a high cross-validation rate and the classification effect.

关 键 词：故障诊断 分类 J48决策树 WEKA 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]