Evolutionary learning of fuzzy neural network using a modified genetic algorithm
Kah Phooi Seng, Kai Ming Tse · Hybrid Intelligent Systems · 2003
This paper presents the tuning of the structure and parameters of a proposed fuzzy neural network (FNN) using a modified genetic algorithm (GA). A FNN with switches introduced to layer 2-3 and 3-4 links is proposed. By doing this, the proposed FNN can learn both the input-output relationships of an application and the network structure using the modified GA. The number of hidden nodes in layer 3 is chosen manually by increasing it from a small number until the learning performance in terms of fitness value is good enough. An application example on sunspot forecasting is given to highlight the merits of the modified GA and the proposed FNN.