主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

Chinese Journal of Management Science ›› 2005, Vol. ›› Issue (1): 30-36.

Previous Articles     Next Articles

A Long-Term Trend Forecasting Approach for Oil Price Based on Wavelet Analysis

LIANG Qiang, FAN Ying, WEI Yi-ming   

  1. Institute of Policy and Management, Chinese Academy of Sciences, Beijing 100080, China
  • Received:2004-10-10 Online:2005-02-28 Published:2012-03-07

Abstract: This paper applies the wavelet method to the oil price long-term trend forecasing.By using the function of wavelet multi-scale analysi,we propose an approach which can accurately predict the future long-term trend of oil price according to the oil price time series.The advantage of the wavelet long-term trend forecast approach is that it can abstract the long-term trend of the oil price accurately and realize the non-linear characteristic of the oil price movements.Thus depending on the historical time series of the oil price,we can figure out the long-term multi-step forecast in a long future.The empirical research is constructed for an one-year long-term trend forecasing of the Brent oil price.By the comparison between the forecast result of this approach with the those of some other time series prediction approaches such as ARIMA,GARCH,Holt-Winters,we demonstrate that the predicted power of the wavelet long-term trend forecast approach in the oil price long-term trend prediction is much better than many other time series forecasting approaches.

Key words: wavelet analysis, oil price, long-term trend, time series, multi-step forecast

CLC Number: