Genetic algorithms for structural optimisation, dynamic adaptation and automated design of fuzzy neural networks
Nikola Kirilov Kasabov, Michael John Watts · Proceedings of International Conference on Neural Networks (ICNN'97) · 2002
Fuzzy neural networks have features which make them useful for knowledge engineering, namely: fast learning; good generalisation; good explanation facilities in the form of fuzzy rules; abilities to accommodate both data and existing fuzzy knowledge about the problem under consideration. This paper presents a current project on using genetic algorithms for optimisation of the structure of a fuzzy neural network called FuNN, for finding the best adaptation mode and for its automated design. Experiments on speech data are reported as part of the project which is aimed at building adaptive speech recognition systems.