A chaotic time series prediction based on neural network: Evidence from the Shanghai Composite index in China

Abstract

The paper compares the performances between the back propagation (BP) neural network and the radial basis function (RBF) neural network in chaotic time series prediction with the Logistic equation, and the results show that the RBF neural network is better than the BP neural network. Further we apply the RBF neural network to predict the Shanghai Composite index that is chaotic according to the phase diagram analysis. The paper reaches the conclusion that it is difficult to predict a chaotic time series over a long period due to the sensitive dependence on initial conditions, but it is feasible to predict a chaotic time series over a short period.