针对大规模互联网搜索数据的多维特性,本文引入动态因子模型思想,构建反映更多关键特征的搜索指数,并使用生物信息学中因果分析方法对指数进行评价.将该指数构建及评价方法应用到国际原油现货和期货价格的分析中,与已有研究相比,本文构造的指数更能全面反映原油市场动态.实证结果表明:本文所构造的搜索指数与原油现货及期货价格均存在显著的因果关系,并能准确探测原油价格的拐点.该方法不仅可用于对国际油价的分析,也易于扩展到其他领域价格分析及预测中.
Due to the multi-dimensional nature of large-scale internet search data,the thesis introduced dynamic factor model to build the search index reflecting more key features.Causal analysis in bioinformatics is also employed to evaluate the search index.Moreover,the concept and evaluation methodology is used in the analysis of spot and futures prices of international crude oil.The empirical study shows that this new search index has a significant causality with both the spot and futures prices of international crude oil,and has superiority in the forecast of the turn point of oil price.This method not only can be used in the analysis of oil price,but also can be expanded to other research fields.