摘要
鉴于WTI原油现货价格序列是一个具有长记忆性的非线性系统,将分数阶差分与非参数自回归模型相结合建立了WTI原油现货价格序列的基于分数阶差分的非参数自回归预测模型,并将该模型与非参数自回归模型和自回归滑动平均模型进行比较.实证研究结果表明:基于分数阶差分的非参数自回归模型的预测精度较高,对WTI原油现货价格的预测较准确.
Considering that WTI crude oil spot price series is a nonlinear system with a long mem- ory, we attempt to frame the nonparametric autoregressive prediction model of the WTI crude oil spot price series based on the fractional differential and nonparametric autoregressive model. Fur- thermore, we compare the new model with non-parametric autoregressive model and autoregres- sive moving average model, respectively. The empirical results show that the nonparametric au- toregressive model based on fractional differential has higher predicted accuracy, and can forecast WTI crude oil spot prices more accurately.
出处
《山东理工大学学报（自然科学版）》
CAS
2012年第6期69-73,共5页
Journal of Shandong University of Technology:Natural Science Edition
关键词
WTI原油现货价格
长记忆性
分数阶差分
非参数自回归模型
WTI crude oil spot price
long memory
fractional differential
nonparametric autore-gressive model
