基于HP滤波—AR模型—GARCH族模型对黄金价格预测研究

Research and prediction of gold price based on the HP filter-AR model-GARCH family model

黄金作为一种特殊的贵金属,不仅其本身具有货币和商品的双重功能,而且对经济领域有着重要影响,因此预测黄金价格趋势对社会经济发展具有重要意义。文中提出了一种新的预测方法:首先采用HP滤波将时间序列分解为趋势要素序列和周期波动序列;然后针对不同序列的性质,对趋势要素序列采用自回归模型(AR)拟合预测,对周期波动序列采用ARMA-GARCH族模型拟合预测;最后将两个预测序列相加与原序列比较;预测结果在模型精度和范围上均令人满意。

Gold as a special kind of precious metal ,not only has the dual function of money and goods ,but also has important influence on economy ,therefore,prediction of gold price trends for investors and the development of so-cial economy are significant .This paper proposes a new forecasting model that first of all decomposes time sequences into trend factor sequences and periodic fluctuation sequences by HP filter .And for different time sequences , fitting prediction of trend factor sequence is made using autoregressive model （AR） model,while fitting prediction of periodic fluctuation sequences is made using ARMA -GARCH family model .Finally those two forecast sequences are added to-gether before compared to the original sequence ,proving to be acceptable on model precision and forecast range .