Application of Improved Multi-classification SVM on Fault Diagnosis for Analog Circuits
For the incipient faults of analog circuit can not be distinguished well by traditional multi-classification SVM, a new SVM fault diagnosis model based on separability measure binary tree (SMBT) was presented. Firstly, the dimensions of the experimental samples were decreased and classified briefly by principal components analysis (PCA); Secondly, best parameters of SVM were found in the sense of cross-validation by genetic algorithms (GA),than the binary tree SVM model was built and trained based on separability measure among samples. Finally, from experimental results, the conclusion can be drawn that SMBT has certain advantages for the incipient fault diagnosis comparing with model of one-against-one (OAO) and one-against-rest (OAR).
SVM SMBT analog circuit incipient fault diagnosis
Renyang LIU Wenquan WU Chao LI Long MA
College of Electronic Engineering Naval University of Engineering Wuhan, China Department of Basic Level Service Military Economy Academy Xiangyang, China
国际会议
三峡
英文
2190-2193
2012-05-18（万方平台首次上网日期，不代表论文的发表时间）