本文選題：原油價格預測 + SVM��； 參考：《大連海事大學》2012年碩士論文

【摘要】：原油作為國際大宗商品,是當今世界最為重要的基礎能源,對一個國家政治的穩(wěn)定和經(jīng)濟的發(fā)展至關重要,原油作為一種戰(zhàn)略物資,往往成為大國博弈的焦點。國際原油的價格由于受突發(fā)事件的影響,常常偏離了正常的市場價格水平,原油價格預測十分困難。原油價格預測問題是關系到原油生產(chǎn)企業(yè)、原油消費企業(yè)和國家利益的重大問題,是國內(nèi)企業(yè)參與國際競爭的重要支撐,原油價格的巨幅波動對國家的政治安全和經(jīng)濟安全產(chǎn)生很大的沖擊,因此能否精確的預測原油價格顯得尤為重要。本文首先從供需角度分析了國際原油的供給與需求對石油價格波動的影響；接著分析了歷年發(fā)生的地緣政治等突發(fā)事件對原油價格的影響,并對其進行了結構斷點檢測,發(fā)現(xiàn)在那些突發(fā)事件發(fā)生的月份,原油價格都會明顯的偏離正常的市場價格,因此可以認為這些突發(fā)事件影響下的原油價格不是反映真實的市場的價格,有理由認定為奇異樣本；在此基礎上,運用SVM對原油價格指數(shù)進行預測,利用遺傳算法對SVM中的核函數(shù)g、懲罰參數(shù)c和損失函數(shù)epsilon值p進行優(yōu)化,然后對預測得到的結果進行ARMA修正,盡可能消除由于信息誤差對于預測結果的影響；實證分析表明,基于模糊粒子GA-SVM-ARMA模型能夠很好的預測原油價格。文章最后還分析了原油價格指數(shù)波動對航運業(yè)的影響,發(fā)現(xiàn)原油價格通過航運成本、航運需求、航運效率三個方面來影響航運市場；為了進一步的分析這種影響,建立了原油價格指數(shù)與BDI指數(shù)的VAR模型,用沖擊響應函數(shù)和方差分解方法得到了油價波動對BDI的沖擊以及貢獻度,結果表明在航運需求周期性不明顯的時間段里,油價對BDI的影響是比較顯著的。
[Abstract]:Crude oil, as an international commodity, is the most important basic energy in the world, which is very important to the political stability and economic development of a country. As a strategic material, crude oil often becomes the focus of the game between big countries. The price of international crude oil is often deviated from the normal market price level because of the sudden events, so it is very difficult to predict the price of crude oil. The prediction of crude oil prices is a major issue related to the interests of crude oil production enterprises, crude oil consuming enterprises and the state. It is also an important support for domestic enterprises to participate in international competition. The huge fluctuation of crude oil price has a great impact on the political and economic security of the country, so it is very important to predict the crude oil price accurately. This paper first analyzes the impact of international crude oil supply and demand on oil price fluctuation from the point of view of supply and demand, and then analyzes the impact on crude oil price caused by unexpected events such as geopolitical events over the years, and carries out structural breakpoints detection. It is found that crude oil prices deviate significantly from normal market prices in the months in which unexpected events occur, so it can be considered that the crude oil prices under the influence of these emergencies do not reflect the true market prices. On this basis, SVM is used to predict the crude oil price index, and genetic algorithm is used to optimize the kernel function g, the penalty parameter c and the epsilon value p of the loss function in SVM. Then the predicted results are modified by ARMA to eliminate the influence of the information error on the prediction results as far as possible. The empirical analysis shows that the oil price can be predicted well based on the fuzzy particle GA-SVM-ARMA model. Finally, the paper analyzes the impact of the fluctuation of crude oil price index on shipping industry, and finds that crude oil price affects the shipping market through three aspects: shipping cost, shipping demand and shipping efficiency. The VAR model of crude oil price index and BDI index is established, and the impact and contribution of oil price fluctuation to BDI are obtained by means of shock response function and variance decomposition method. The results show that the periodicity of shipping demand is not obvious in the time period. The effect of oil price on BDI is significant.
【學位授予單位】：大連海事大學
【學位級別】：碩士
【學位授予年份】：2012
【分類號】：F416.22;F551;F224

【參考文獻】

相關期刊論文 前6條

1 常松,何建敏;基于小波包和神經(jīng)網(wǎng)絡的股票價格預測模型[J];中國管理科學;2001年05期

2 林敏;;社會政治因素對石油價格影響的定量化描述方法[J];中國石油大學學報(社會科學版);2007年02期

3 梁強,范英,魏一鳴;基于小波分析的石油價格長期趨勢預測方法及其實證研究[J];中國管理科學;2005年01期

4 金晶;王行愚;羅先國;王蓓;;PSO-ε-SVM的回歸算法[J];華東理工大學學報;2006年07期

5 杜樹新,吳鐵軍;用于回歸估計的支持向量機方法[J];系統(tǒng)仿真學報;2003年11期

6 赫叢喜;安亞娜;王振喜;;國際化石油價格形成機制研究[J];遼寧工程技術大學學報(社會科學版);2006年02期

，

本文編號：1903451