A new fuzzy membership function for FSVM and its application in machinery fault diagnosis

In this paper, a new fuzzy membership function for fuzzy support vector machine is presented. It provides an effective approach to deal with the over-fitting problem when outliers exist in the training data set. Combining with the concept of the K-nearest neighbor algorithm, we give a definition of the new fuzzy membership function. Then, fuzzy support vector machine with some improvements is successfully applied in machinery fault diagnosis and some engineering experimental results show the good performance of the present approach.