Parameter optimization for Takagi-Sugeno fuzzy models-lessons learnt

Describes an approach to automatically build a Takagi-Sugeno fuzzy model (TSK-model) based on a set of input-output data. Identifying rule-based fuzzy models consists of two parts: structure modeling and parameter optimization. For structure modeling, we investigate several search heuristics. In order to find a good model structure, such search heuristics make it necessary to optimize and then evaluate a large number of different candidate models. To be applicable to real world problems, the parameter optimization must be highly efficient. For this, we investigate the use of two gradient descent algorithms: standard gradient descent (backpropagation) and resilient propagation (RPROP). The combination of a structure search with a fast parameter optimization yields a powerful modeling algorithm that is capable to identify large real world systems. We evaluate a number of varieties of TSK-like fuzzy models by performing a nonlinear regression benchmark of Frank (1995). We compare several types of fuzzy models that are constructed as combinations of different fuzzy sets, structuring algorithms, and parameter optimization techniques