摘要
油价预测是能源市场研究的一个重要领域.本文基于季节调整技术和周期性分析技术,提出了一个新的预测模型-季节-谐波模型,其思路是分解-组合.通过对五种油品(WTI原油、Brent原油、无铅汽油、柴油和取暖油)价格的实证分析,与其它五种方法(ARIMA模型、指数平滑、Winters方法、EGARCH模型和逐步自回归)相比,季节-谐波模型取得了最好的预测效果.
The oil price forecasting is an important field in energy market research. The paper establishes a new model - the season-harmonic model based on the seasonally adjustment technology and cyclical analysis, its idea is decomposition-combination. To five oil prices （WTI crude oil, Brent crude oil, unleaded gasoline, diesel and heating oil）, empirical analysis shows that season-harmonic model obtains the best forecasting results compare to the other five methods （ARIMA model, exponential smoothing, Winters method, EGARCH model and step by step autoregression model）.
出处
《数理统计与管理》
CSSCI
北大核心
2009年第3期395-401,共7页
Journal of Applied Statistics and Management
基金
北京市优秀人才培养资助项目(20071D0500200137)
教育部人文社会科学研究青年基金项目(08JC790004)
关键词
季节调整
周期性分析
油价预测
季节-谐波模型
seasonally adjustment, periodicity analysis, oil price forecasting, season-harmonic model
