Boya Century Publishing

Set Homepage | Favorite
Article Details


Article Title : ARIMA-SVR Decomposition Integrated Crude Oil Price Forecasting Model based on X12-HP Filter
Author(s) : Yufeng Yang
Corresponding Author : Yufeng Yang
Keywords : X12 Seasonal Separation; HP Filter; Crude Oil Price Forecast; ARIMA; SVR.

In view of the importance of crude oil price in the world economy, accurate crude oil price forecast has been widely concerned. The crude oil price has the characteristics of non-linear and high noise, so it is a complex and challenging work to accurately predict. First, this paper expounds on the structure of China's crude oil production and consumption, which not only shows that China's crude oil is too dependent on foreign countries but also shows that the fluctuation of international crude oil will bring a lot of uncertain impact to China's economy. Therefore, it is necessary to forecast the international oil price. To solve this problem, this paper proposes a hybrid model to forecast crude oil price based on the X12-HP filter, which is called ARIMA-SVR. The seasonal factors of the oil price are eliminated by Census-X12, the crude oil price series after seasonal elimination is decomposed by HP filter, and the long-term trend price and periodic fluctuation price of crude oil are predicted by ARIMA and SVR models respectively. Finally, we integrate the two-forecasting series to get the final crude oil price forecast. We give an empirical analysis based on the WTI crude oil price data from January 2002 to March 2021. We divide the data into in-sample data and out-of-sample data. The empirical results show that the proposed integrated model can significantly improve the prediction accuracy of WTI crude oil price compared with the traditional ARIMA and GARCH models. Finally, we discuss the sustainable development of energy and give relevant policy suggestions.