基于非线性组合方法的石油期货价格预测研究  

Petroleum Futures Price Prediction Research Based on Nonlinear Combination Method

作　　者：张凯[1] 沙锋[1] 

机构地区：[1]河南城建学院,河南平顶山467036

出　　处：《计算机仿真》2012年第7期379-382,共4页Computer Simulation

摘　　要：研究石油期货价格预测问题,石油期货价格是一种动态、时变、非线性数据,具有较强的随机性和多元性。单一预测模型只能描述部分信息,存在局限性。为了提高石油期货价格预测精度,提出了一种非线性组合的石油期货价格预测模型。首先采用3个不同的单项模型对石油期货价格进行预测,采用全局逼近能力强的支持向量机建立了一个多输入单输出的石油期货价格非线性组合预测模型,并通过预测误差最小原则确定模型的参数。仿真结果表明,改进模型预测精度明显优于对比预测模型,为提高石油期货价格预测精度提供了一条新途径。The price of oil futures is a kind of dynamic, time-varying, nonlinear data, single forecasting model can only describe the partial information. In order to improve the petroleum futures price prediction accuracy, this pa- per presented a petroleum price prediction model based on nonlinear combination. Firstly, three different petroleum price prediction models werer used, then the universal approximation ability of support vector machine was used to establish a nonlinear combination forecast model, and through the principle of minimum prediction error, the parameters of the model was determined. The simulation results show that, the model prediction accuracy is obviously better than the contrast model, ,cam improve the petroleum futures price prediction accuracy.

关 键 词：石油期货价格 组合预测 支持向量机 

分 类 号：F224[经济管理—国民经济]