首页 | 本学科首页   官方微博 | 高级检索  
     检索      

结合交叠区异点统计和相关分析的免疫克隆特征选择方法
引用本文:栗茂林,梁霖,王孙安,刘弹.结合交叠区异点统计和相关分析的免疫克隆特征选择方法[J].西安交通大学学报,2012,46(5):50-56.
作者姓名:栗茂林  梁霖  王孙安  刘弹
作者单位:西安交通大学机械工程学院 710049 西安
基金项目:国家自然科学基金资助项目(51075323);中央高校基本科研业务费专项资金资助项目(xjj20100066)
摘    要:为了解决机械故障诊断中的特征选择问题,利用免疫克隆选择算法,提出了一种结合交叠区异点统计和相关性分析的免疫克隆特征选择方法,可有效地去除不相关特征和冗余特征.基于空间分布的交叠区异点,设计了交叠区异点统计的优化指标;基于J散度距离,设计了一种冗余特征的评估指标;基于免疫克隆选择算法,结合交叠区异点统计优化指标和冗余特征评估指标,提出了一种免疫克隆特征选择新算法.仿真和实际工程应用的结果表明:文中提出的方法比常用的特征选择方法更加有效,选出的特征分类精度更高,特征子集更小,更能满足故障诊断的需求.

关 键 词:故障诊断  特征选择  免疫克隆选择算法  相关分析

Immune Clone Feature Selection Method Combining Different Point Statistics of Overlap Region with Correlation Analysis
LI Maolin , LIANG Lin , WANG Sun'an , LIU Dan.Immune Clone Feature Selection Method Combining Different Point Statistics of Overlap Region with Correlation Analysis[J].Journal of Xi'an Jiaotong University,2012,46(5):50-56.
Authors:LI Maolin  LIANG Lin  WANG Sun'an  LIU Dan
Institution:(School of Mechanical Engineering,Xi’an Jiaotong University,Xi’an 710049,China)
Abstract:To treat the feature selection for mechanical fault diagnosis,a novel immune clone feature selection method combining different point statistics of overlap region with correlation analysis is proposed with which the irrelevant features and redundant features can be removed effectively.According to the different point of the overlap region in space distribution,an optimization index is designed.An evaluation index of redundant features in correlation analysis following J-divergence is put forward.Then an immune clone feature selection strategy(ICFSS) is constructed with the two indexes.The simulation and industrial applications show its higher precision and smaller feature subset for more effective fault diagnosis.
Keywords:fault diagnosis  feature selection  immune clone selection method  correlation analysis
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号