Oil Futures Prices And Spot Prices

For the forecasting of oil futures prices, this article will adopt "divide and conquer" thought complex oil futures prices prediction problem divided into three relatively simple sub-problems to research. The first step, with clustering analysis idea, divided into several categories target samples; The second step, use discriminant analysis thoughts, on the basis of the previous step discrimination observed quantity and treat a judgment, determine its both belong categories; Finally, according to the ownership of the paper maintains by measuring different categories, we adopt different support vector machine return to describe various other input/output relationship. That is to say, in oil futures prices forecast, subordinate to the input data to discrimination, then select categories with corresponding category of support vector machine (SVM) model forecast. Through an empirical study found that this method are better able to extract data contains information, reflect the evolution regularity of the data, so as to increase the prediction accuracy and effectiveness.In addition the paper oil spot price and oil futures prices on the futures price, the relationship with the spot price establishment relevant analysis model, and then build between both regression analysis models of spot price, forecast. Through an empirical research found that the futures price and spot price is highly related, and use the multivariate linear regression analysis method to establish the linear regression model, use nonlinear regression analysis method to establish the nonlinear regression model. After the above model again to establish inspection result shows that the regression model has very significant statistical significance, for oil spot price forecast has very good effect.