Research of Wavelet Neural Network Model Based on Extenics

摘要

In order to conquer the disadvantage of the traditional Wavelet Neural Networks (WNN), the paper presents WNN model which is based on extenics. The model uses the feature number of matter element to determine the volume of importation of neuron number, makes sure the number of output neurons based on identification the type number needed, and makes a correct judgment about the neurons inhibit or activation. It optimizes the structure design of wavelet neural network. Then there is an experiment about the weather prediction using the new model. Experimental results show that the new model has better convergence and accuracy.