Multi-Class Semi-Supervised Learning in Machine Condition Monitoring

Currently, condition-based maintenance becomes more and more important with the addition of factory automation through the development of new technologies. For many complicated machines, it is difficult to use the mathematical model to describe their faults. Intelligent maintenance makes it possible to perform maintenance similar to that of a human being. Support vector machine (SVM) has become famous and popular in artificial intelligent community due to the excellence of generalization ability than the traditional method such as neural network. However, conventional SVMs use only labeled data (feature/label pairs) to train. Labeled instances are often difficult, expensive, or time consuming to obtain. Semi-supervised learning addresses this problem by using large amount of unlabeled data, together with the labeled data, to build better classifiers. In this paper, a multi-class semi-supervised framework is proposed to perform classification for machine condition monitoring. The effectiveness of the framework is verified by the application to the bearing diagnosis.