A novel hybrid method based on kernel-free support vector regression for stock indices and price forecasting

Abstract

Price forecasting in the financial market is one of the most important and challenging tasks in the field of time series forecasting since it is noisy, non-linear and non-stationary. In this paper, we first develop a kernel-free support vector regression model which not only has a strong flexibility to capture the nonlinear structure of the data but also maintains the high efficiency to avoid choosing a suitable kernel and its related parameters. Then a novel hybrid method is proposed combining empirical mode decomposition algorithm, quadratic surface support vector regression and autoregressive integrated moving average method for the stock indices and future price forecasting. This ensemble scheme fully takes the advantages of these individual methods to efficiently produce accurate time series forecasts. Finally, to compare our proposed method with other benchmark forecasting methods, three stock indices and three future prices are selected as the forecasting targets. The numerical results and statistical test strongly demonstrate the promising performance of our proposed hybrid method in terms of forecasting accuracy, efficiency and robustness.

Keywords:

• Autoregressive integrated moving average
• empirical mode decomposition
• financial time series forecasting
• kernel-free support vector regression
• quadratic surface

Acknowledgments

The authors are listed in reverse alphabetical order.

Additional information

Funding

Luo’s research has been supported by the National Natural Science Foundation of China Grant #72261008. Tian’s research has been supported by the Fundamental Research Funds for the Central Universities #JBK2203005.