Modeling via on-line clustering and fuzzy support vector machines for nonlinear system
Julio César Tovar, Wen Yu, Floriberto Ortiz, Carlos Roman Mariaca, José de Jesús Rubio · 2011
This paper describes a novel non-linear modeling approach by on-line clustering, fuzzy rules and fuzzy support vector machines. Structure identification is realized by on-line clustering method and support vector machines, and the rules are generated automatically. Time-varying learning rates are applied for updating the membership functions of the fuzzy rules. Finally, tue upper bounds of modeling errors are proven.