Hybrid identification in fuzzy-neural networks

✍ Scribed by Sung-Kwun Oh; Witold Pedrycz; Ho-Sung Park

• Publisher: Elsevier Science
• Year: 2003
• Tongue: English
• Weight: 600 KB
• Volume: 138
• Category: Article
• ISSN: 0165-0114
• DOI: 10.1016/s0165-0114(02)00441-4

✦ Synopsis

This paper introduces an identiÿcation method for nonlinear models in the form of Fuzzy-Neural Networks (FNN). In this model, we use two forms of the fuzzy inference methods-a simpliÿed and linear fuzzy inference, and exploit a standard Error Back Propagation learning algorithm. The FNN modeling and identiÿcation environment realizes parameter identiÿcation through a synergistic usage of clustering techniques, genetic optimization and a complex search method. We use a Hard C-Means (HCM) clustering algorithm to determine initial apexes of the membership functions of the information granules used in this fuzzy model. The parameters such as apexes of membership functions, learning rates, and momentum coe cients are then adjusted using hybrid algorithm. The proposed hybrid identiÿcation algorithm is carried out by combining both genetic optimization (genetic algorithm, GA) and the improved complex method. An aggregate objective function (performance index) with a weighting factor is introduced to achieve a sound balance between approximation and generalization of the model. According to the selection and adjustment of the weighting factor of this objective function, we reveal how to design a model with sound approximation and generalization abilities. The proposed model is experimented with using several time series data (gas furnace, sewage treatment process and NO x emission process data of gas turbine power plant).