Font Size: a A A

Combined Non-Linear Forecasting To International Crude Oil Price

Posted on:2007-03-12Degree:MasterType:Thesis
Country:ChinaCandidate:B LiFull Text:PDF
GTID:2189360212480542Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
Petroleum is regarded as one of the most important fundamental resources, industrial chemicals and strategic materials, which supports the economic growth of oil production and consumption countries and is closely concerned with better life and national defense. Oil price fluctuation holds the balance of politics and economy not only in a country, but also in the whole world.Oil price forecasting is of significance to crude oil production and consumption companies and countries, which provides a basis for production and consumption decision. As is known, a right decision brings opportunities and development, which requires accurate information sources for a guarantee. That is why oil price forecasting is attached much importance in a world-wide range.Since June 1, 1998, a new price and circulation mechanism has been carried out in China in line with international oil prices, i.e., domestic oil prices fluctuate according to international oil price fluctuation. This paper focuses on the highly non-linear feature of international oil prices and puts forwards a new forecasting model based on fuzzy neural network, combining RBF neural network, Markov chain based semiparametric model and wavelet analysis based forecasting model with the Brent oil price series. The good results confirm the performance of the model in accurate forecasts, reinforcement learning properties and mapping capabilities.The paper develops as follows. Section 1 introduces crude oil price forecasting research background and significance. Section 2 provides an analytical introduction to correlative theories and models, research objective and thread. Section 3 and 4 are the core of this paper. In section 3, individual forecasting methods and models adopted in this paper are introduced at the objective and non-linear combining forecasting to crude oil price based on fuzzy neural network is put forward. A detailed specification of the forecasting procedure with Brent oil price series is provided in Section 4. Section 5 provides the concluding remarks and a general discussion on the reform of price formation mechanism...
Keywords/Search Tags:crude oil price, RBF network, Markov chain, Semiparametric, wavelet analysis, combined forecasting
PDF Full Text Request
Related items