基于混合模型的原油价格混沌预测方法

针对原油现货价格的非线性和时变性特征,提出一种小波变换结合Elman神经网络和广义自回归条件异方差(GARCH)模型的混沌预测方法。首先利用小波变换将原油现货价格序列分解和重构成概貌序列和细节序列。其次对概貌序列和原油期货价格序列进行相空间重构,建立Elman神经网络的混沌时间序列模型预测概貌序列的未来值;同时以细节序列为历史数据,构建GARCH模型预测细节序列的未来值;最后将概貌序列和细节序列的未来值求和作为最终的预测值。实验证明该方法能够提供更准确的预测结果。

Crude Oil Price Chaotic Forecasting Method Based on Hybrid Model

Due to the characteristics of crude oil spot price, such as non-linear, time-varying, this paper proposes a new chaotic prediction approach based on wavelet transform, Elman neural network and Generalized Autoregressive Conditional Heteroskedastic model. Firstly, the crude oil spot price series is decomposed and reconstructed into approximate series and detailed series. Secondly, the phase space of approximate series and crude oil futures price series are reconstructed. Then, the approximate future values are predicted by chaotic time series model based on Elman neural network. While the detailed future values are forecasted by GARCH model, using the detailed series as historical data. Finally, the sum of the approximate and detailed future values is used as the final forecasting values. The example indicates that the proposed method can provide more accurate forecasted results.