Forecasting Tehran Stock Exchange Index Returns Using a Combination of Wavelet Decomposition and Adaptive Neural Fuzzy Inference Systems

Abstract

In this paper, a framework for time series prediction is presented which makes it possible to predict the future values of a time series more accurately using soft computing approach. In this method, input data of adaptive neural fuzzy inference systems are reduced using wavelet decomposition of random noises; therefore, it reduces errors and improves the desired chaotic time series prediction. The above method was evaluated using Tehran Stock Exchange return series for the period of 23/10/2009 to 23/3/2013, and the results indicate the superiority of the proposed method compared to other ones.

Keywords

• wavelet decomposition
• adaptive neural fuzzy inference systems
• soft computing
• de-noising
• stock exchange