在模式识别系统(PMRS)的基础上,采用石油期货价格加权,提出了能够有效的预测短期石油现货价格的新方法.以自然预测模型(naieve forecast model)和PMRS作为评价基准,对美国西德克萨斯轻质原油(WTI)、普通汽油(regular gasoline)和燃料油(heating oil)的现货价格和期货价格进行了实证研究.结果表明,提出的基于PMRS的期货加权油价多步预测方法具有更为优越的短期预报性能.
In this paper, a futures-weighted short-term oil price prediction approach based on PMRS is proposed. Empirical study on the spot and futures price of WTI, regular gasoline and heating oil is conducted. The comparison between the new approach and the naieve forecast model as well as the PMRS model indicates that the futures-weighted oil price multi-step prediction approach based on PMRS performs much better for short-term oil price prediction.