Fuzzy modelling via on-line support vector machines

Abstract

This article introduces an approach to identify unknown nonlinear systems by fuzzy rules and support vector machines (SVMs). Structure identification is realised by an on-line SVM technique, the fuzzy rules are generated automatically. Time-varying learning rates are applied for updating the membership functions of the fuzzy rules. Finally, the upper bounds of the modelling errors are proven.

Keywords:

• fuzzy logic
• support vector machines
• identification
• clustering