Application of Twin Support Vector Machine for fault diagnosis of rolling bearing

Abstract

The number of fault samples is only the small portion in the whole sample set. How to diagnose the rolling bearing fault accurately becomes a challenge in the unbalance sample set. Twin Support Vector machine (TWSVM) is applied into the bearing fault diagnosis in the study. It aims at generating two nonparallel planes in which each plane is closer to one of the two classes and is as far as possible from the other. The fault diagnosis experiments verify that TWSVM has higher accuracy and faster speed than Support Vector Machine, and identify the bearing fault well. © Springer International Publishing Switzerland 2014.