A new hybrid approach for forecasting of daily stock market time series data


Abstract


In recent years, many researchers have focused on forecasting financial time series data, especially stock market data. Stock market data possesses so many features that forecasting may be very challenging. In the present study, a hybrid of two methodologies is proposed, which is the Empirical Mode Decomposition (EMD) and the Random Walk (RW) in order to enhance the stock market forecasting performance, denoted by (EMD-RW). The advantage of EMD-RW is its ability to forecast nonlinear and nonstationary stock market data without the need to use some transformation method or differencing a time series technique. Moreover, the new proposed EMD-RW produced high-accuracy results. Ten stock market time series for ten different countries are used in this study to demonstrate the forecasting accuracy of the EMD-RW. Results using four forecasting accuracy functions display that EMD-RW forecasting accuracy is better than the four compared methods.

DOI Code: 10.1285/i20705948v17n1p162

Keywords: EMD, Forecasting, Nonstationary time series

Full Text: pdf
کاغذ a4

Creative Commons License
This work is licensed under a Creative Commons Attribuzione - Non commerciale - Non opere derivate 3.0 Italia License.