Transactions of China Electrotechnical Society  2016, Vol. 31 Issue (4): 28-35    DOI:
Orginal Article Current Issue| Next Issue| Archive| Adv Search |
Analog Circuit Fault Diagnosis Based on Local Graph Embedding Weighted-Penalty SVM
Liao Jian1, 2, Shi Xianjun2, Zhou Shaolei2, Xiao Zhicai2
1. The Troop 91550 of PLA Dalian 116000 China;
2. Department of Control Engineering Naval Aeronautical and Astronautical University Yantai 264001 China

Download: PDF (499 KB)   HTML (1 KB) 
Export: BibTeX | EndNote (RIS)      
Abstract  This paper proposes a method for analog IC diagnosis based on local graph embedding weighted-penalty support vector machine (SVM), to overcome the shortcomings of fault diagnosis method based on traditional SVM. A new type of SVM lying on data distribution is designed. Herein, maximize the inter-class margin of the entire data while optimizing local distribution of the data manifold, meanwhile, introduce the global data distribution information in error costs. The proposed method effectively combines the prior distribution information to improve the robustness and increase the diagnosis accuracy. The simulation results show the effectiveness of the algorithm.
Key wordsAnalog circuit      fault diagnosis      support vector machine      data manifold     
Published: 03 March 2016
PACS: TP181  
  TM930  
Fund:国家自然科学基金重点资助项目(61203168)
Corresponding Authors: 廖 剑 男,1985年生,博士,主要研究方向为电路设计、测试与诊断。E-mail: 251250544@qq.com   
Service
E-mail this article
Add to my bookshelf
Add to citation manager
E-mail Alert
RSS
Articles by authors
Liao Jian
Shi Xianjun
Zhou Shaolei
Xiao Zhicai
Cite this article:   
Liao Jian,Shi Xianjun,Zhou Shaolei等. Analog Circuit Fault Diagnosis Based on Local Graph Embedding Weighted-Penalty SVM[J]. Transactions of China Electrotechnical Society, 2016, 31(4): 28-35.
URL:  
https://dgjsxb.ces-transaction.com/EN/     OR     https://dgjsxb.ces-transaction.com/EN/Y2016/V31/I4/28
Copyright © Transactions of China Electrotechnical Society
Supported by: Beijing Magtech