摘要
在考察前人油价预测模型研究情况后,文章一方面,从油价序列长记忆性和异方差性着手,将ARFIMA模型和GARCH模型进行结合,构建ARFIMA-GARCH模型;另一方面,还对油价模型构建中的一大难题——影响因素的筛选进行适当探索,尝试结合主成分分析,提取若干主成分,加入ARFIMA-GARCH模型中,形成基于PCA的ARFIMA-GARCH模型。在与其他模型进行比较好,发现基于PCA的ARFIMA-GARCH模型要好于其他模型,文章的研究和改进是有效的和成功的。
On the consideration of the forecasting model for oil price that others had study, the article, on the one hand took long memory and heteroskedasticity into account, building ARFIMA-GARCH model by combining ARFIMA and GARCH; on the other hand, explored properly the filtering of the factors ,which is one of the hardest problem for forecasting. Compared to other models, the ARFIMA-GARCH model based on PCA performed better. The study and improvement was effective and successful.
出处
《价值工程》
2011年第27期102-104,共3页
Value Engineering
关键词
主成分分析
广义自回归条件异方差模型
分整自回归移动平均模型
principal components analysis
generalized autoregressive conditional heteroskedastic model
autoregressive fractional integrated movingaverage model