Activation Function of Wavelet Chaotic Neural Networks

Abstract

In this paper, we studied the activation function of chaotic neural network, at first we review Chen's chaotic neural network and then propose a novel chaotic neural network model whose activation function is compose of sigmoid and wavelet function. Second, we make an analysis of the largest Lyapunov exponents and the reversed bifurcations of the neural units of Chen's and the proposed model. Third, 10-city traveling salesman problem (TSP) is given to make a comparison between them. Finally we conclude that the novel chaotic neural network model we propose is more valid