Font Size: a A A

Study On The ARFIMA-GARCH Model For Forecasting Oil Price Based On PCA

Posted on:2013-01-30Degree:MasterType:Thesis
Country:ChinaCandidate:W C WangFull Text:PDF
GTID:2219330362961386Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
After the twenty-first century, the fluctuation of international crude oil price became more complicated, and the regular pattern of fluctuation became more difficult to identify. Modeling for international crude oil market is more difficult. However, in the meanwhile, the necessity of the modeling has been reinforced. With the rapid growth of economy, the demand for crude oil has been more and more intensive. Gradually, two problems are manifest: on the one hand, the difficulty of prediction has made the petrochemistry enterprise and the transportation industry threaten seriously; on the other hand, because of the high degree of dependence on crude oil import, the fluctuation not only jeopardized the national economy, but also have a far-reaching influence in terms of geopolitics and the right to speak. Many kinds of factors and events made the situation being desperate. Again, the tendency of crude oil price became the important issue for the experts and researchers.At that point, this article that took the forecasting of international crude oil price as research object, firstly made a description about the impact that the fluctuation of crude oil price had on our country's economy; secondly, this article reviewed the history of research on the international crude oil price from both factors research and prediction model research. On the existed researches, there was limitation in the selecting of variables, but the causality in real world was complicated. The representativeness of the factors selected was doubtful. And the time series model in existed researches only took the feature of the dependent variables own into consideration, rather than contained exogenous variables. So the information contained in the model was not enough. Basing on the analysis, this article make a principle component analysis on the data pool which contained 16 variables to reduce dimensions. After that, by combining the principle factors and time series modeling, this article build a hybrid model: the ARFIMA-GARCH model based on the PCA. The results proved, the mixture of PCA, ARFIMA and GARCH, has made the prediction more effective and accurate.
Keywords/Search Tags:international crude oil price, principal components analysis, autoregressive fractional integrated moving average model, generalized autoregressive conditional heteroskedastic model
PDF Full Text Request
Related items