Improved prediction of financial market cycles with artificial neural network and markov regime switching

Abstract

This paper provides an analysis of the Shanghai Stock Exchange Composite Index Movement Forecasting for the period 1999-2009 using two competing non-linear models, univariate Markov Regime Switching model and Artificial Neural Network Model (RBF). The experiment shows that RBF is a useful method for forecasting the regime duration of the Moving Trends of Stock Composite Index. The framework employed also proves useful for forecasting Stock Composite Index turning points. The empirical results in this paper show that ANN method is preferable to Markov-Switching model to some extent. © 2011 Springer Science+Business Media B.V.