High speed learning of neural network using fuzzy logic

We have proposed a fuzzy method to vary and modify the critical coefficients ETA and ALPHA involved with the feed forward multilayer neural network (FF network) to raise the network learning speed. In fact a fuzzy controller has been designed in this regard in order to accept the absolute and the change of error values as its inputs to determine the proper ETA and ALPHA coefficients. This can be done with the aid of the appropriate rule bases that result from the observation and analysis of neural network behavior during its learning interval