Research of fault diagnosis based on rough sets and support vector machine
Du Anli, Yingchun Wang, Jie Wang, Hua Jiajun, Shao Mengguo · 2011
It is lack of fault samples and the feature information is miscellaneous and redundant in complex circuit system. In order to solve the problem, a new fault diagnosis method was presented based on rough set (RS) and support vector machine (SVM). The RS was applied to discrete sample data the genetic algorithm (GA) was used to reduce the redundant attributes and the conflicting samples. Then the simplest fault attributes were extracted as the training samples for SVM, which was used as the classifier to isolate the faults rapidly. The simulated experiments demonstrated that the method is valid and feasible under the condition of small samples.