Author(s)

Wang Pingan, Gan Xusheng, Li Huaping

Corresponding Author

Wang Pingan

Abstract

For the parameter initialization and network structure determination of Wavelet Neural Network (WNN) modeling, an optimization design method based on modified Artificial Fish Swarm Algorithm (AFSA) is proposed. Firstly, the effect of initial parameters and network structure on the performance of WNN model is discussed, and then an AFSA with different fish foraging behavior is applied to determine the initial parameters and the number of hidden nodes required in modeling. The experiment shows that, the proposed method can completely solve the optimization design problem in WNN modeling.

Keywords

RBF Neural Network, Artificial Fish Swarm Algorithm, Stock Price Trend, Prediction.