单一的线性模型或单一的非线性模型都难以全面反映黄金价格波动性的全部信息,根据单整自回归移动平均(ARIMA)模型和神经网络(NN)模型的各自特点,以上海黄金交易所黄金现货市场的交易品种Au99.99为研究对象,建立了ARIMA模型融合神经网络的黄金价格时间序列预测模型。实证结果表明,融合模型ARIMA-NN的预测准确性明显比单一模型的准确性高。
It is difficult to reflect all information of the gold price by single linear or nonlinear models. Based on respective characteristics of the Autoregressive Integrated Moving Average(ARIMA) and Neural Networks(NN) models, a gold price time series forecasting model combining the two is established with the trading product Au99.99 in the Shanghai gold exchange as the research object. The empirical results demonstrate that the combined model ARIMA-NN has a better forecasting accuracy than one single model does.