ABSTRACT

Given the importance of crude oil prices in the world economy, accurate price prediction has drawn extensive attention. Nevertheless, because of the complexity of the crude oil market, most traditional forecasting algorithms fail to meet the accuracy requirements. To achieve higher precision, this paper proposes a novel hybrid model for crude oil price forecasting by combining a Hodrick-Prescott filter with X12 methods and adjusting the order used. Application of our model on both West Texas Intermediate and Brent oil prices forecasting demonstrates its accuracy. The results of various forecasting performance evaluation criteria indicate that the model has stronger stability and better accuracy. The mechanism of seasonal and periodic factors is also analyzed, which provides remarkable references to other time-series predictions. Establishing two different types of predictive models that combine multiple knowledge effectively has obvious advantages over other models and provides more reliable cutting-edge information for designing a Chinese energy development strategy.

摘要

鉴于原油价格在世界经济中的重要性，准确的价格预测已引起广泛关注。尽管如此，由于原油市场的复杂性，大多数传统的预测算法无法满足准确性要求。为了获得更高的精度，本文通过结合Hodrick-Prescott过滤器和X12方法并调整使用的顺序，提出了一种新颖的原油价格混合模型。我们的模型在西德克萨斯中质油和布伦特油价预测中的应用证明了其准确性。各种预测性能评价标准的结果表明，该模型具有较强的稳定性和较好的准确性。还分析了季节性和周期性因素的机制，这为其他时间序列预测提供了出色的参考。