Stock market forecasting using empirical mode decomposition with holt-winter

Abstract

This study aims to implement a hybrid method based on Empirical Mode Decomposition and Holt-Winter (EMD-HW) to predict the stock market. The methodology of this article summarizes as: First, the stock market data are decomposed by EMD method into Intrinsic Mode Functions (IMFs) and residual components. Secondly, all components are forecasted by HW technique. Finally, forecasting values are summed together to get the forecasting value of stock market data. Empirical results showed that the EMD-HW outperform individual forecasting models. The daily stock market time series data for India, Brazil, Indonesia, and Italy are applied to show the forecasting performance of the EMD-HW method. The strength of this EMD-HW lies in its ability to forecast non-stationary and non-linear time series without a need to use any transformation method. Moreover, EMD-HW has a relatively high accuracy comparing with seven existing forecasting methods based on the five forecast error measures.