General fuzzy neural network: theory and applications

We proposed a new general fuzzy neural network (GFNN) which can process the structural self-organizing and parametric self-learning online. Based on the concept of fuzzy logic system, GFNN has a structure of a three-layer feedforward network and a wider information interface. Its neurons in the hidden layer correspond to the fuzzy rules and have some functions of information storage and inference. It is easy for GFNN to realize the novel structural self-organizing online with the presentation of three concepts: the match degree; the match degree criterion; and the minimum believable level. So it can be used to identify the time-varying nonlinear system online.