摘要
由于次贷危机的发生、国际形势的动荡以及投资方面的需要,黄金越来越受到人们的关注,预测黄金价格及其波动性愈益重要。ARIMA时间序列模型虽能较好地把握金融时间序列的动态规律,但不能反映黄金价格的波动特征和杠杆效应。而条件异方差GARCH模型能较好地反映金融时间序列的波动性,更适合金融领域的预测。采用2018年1月2日至2018年12月28日上金所黄金Au(T+D)收盘价日数据,将ARIMA与GARCH相结合,构造新的ARIMA-GARCH族混合模型,并通过ARIMA-GARCH族在不同分布下对黄金价格进行预测,结果发现,黄金价格波动不仅具有异方差性,而且还具有杠杆效应。经过与ARIMA模型对比,发现新的混合模型在一定程度上能更好地拟合黄金价格的运动路径,在短期预测方面更具实用性。
Due to the subprime crisis,the turmoil in the international situation and the need for investment,gold has attracted more and more attention,and it is increasingly important to predict the price of gold and its volatility.ARIMA model can better grasp the dynamic law of financial time series,but it can not reflect the volatility characteristics and leverage effect of gold price.GARCH is more suitable for financial prediction because of its better volatility of financial time series.Based on the closing date data of Au(T+D)gold from January 2,2018 to December 28,2018,this paper combines ARIMA with GARCH effect to construct a new ARMA-GARCH family hybrid model and forecasts the gold price through the ARMA-GARCH family under different distributions(normal distribution,t distribution,generalized error distribution).The results show that gold price fluctuations are not only heteroscedastic but also leveraged.After comparing with the ARIMA model,it is found that the new hybrid model ARMA(4,1,4)-TGARCH(1,1)can better fit the movement path of gold price to a certain extent,and is more practical in short-term prediction.
作者
丁磊
郭万山
DING Lei;GUO Wan-shan(School of Economics,Liaoning University,Shenyang 110036,China;School of Mathematics Science,Harbin Normal University,Harbin 150025,China)
出处
《许昌学院学报》
CAS
2019年第6期124-129,共6页
Journal of Xuchang University
基金
黑龙江省高等学校青年创新人才项目“有关热传导方程的高效数值算法”(UNPYSCT-2017179)
