Study Of Chaotic Time Series Forecasting Model Based On Immune RBF Neural Networks

The chaotic time series forecasting is a process, which we should base the predictive model on the chaotic characteristic of the historical information, and then confirm its parameters in accordance with the historical data, and forecast the target point. The chaotic time series based on the feed-forward neural networks has gradually becomed a very important issue in the chaotic time series research.When the input is a large amount of data, and also is high dimensional, then the RBF neural network would need large memory and also appear local convergence. It uses the global convergence of immune clonal algorithm to find the hidden layer center of the RBF neural network. The weights of the output layer are determined by adopted the least square method. It can improve convergence speed and precision of the RBF neural network. Then, a radial basis function neural network learning algorithm based on immune clonal algorithm is studied. Also, it constructs an chaotic time series forecasting model based on the immune RBF neural network. By the simulation of the chaotic time series, it documents the validity and feasibility of the improved chaotic time series prediction model.