最优波动率模型选择及其在黄金市场中的应用

文章先总结了波动率模型过去的研究，并对不同波动率模型的评估提出三种方法，然后讨论了这些方法在黄金市场波动率预测中的应用。通过分析黄金市场1975年到2004年的数据，得出的结论是，如果基于样本外四期预测误差的评估，EWMA模型较优；如果基于样本外四期预测的R平方的评估，T-GARCH模型较优；如果基于VAR损失函数的真实性检验评估，EWMA模型较优。最后对未来关于金融市场波动率的研究提出一些建议。

Choice of Best Volatility Models and Its Application to Gold Market

For giving an accurate forecast for financial market, the article discusses the volatility and its evaluation. Firstly it summarizes the achievement of the research, and investigates some methods for choosing the volatility models. Then it discusses its application to gold market. By analyzing the data of gold market from 1975 to 2004, we conclude that, if based on the four weeks of out-of-sample error forecast, the EWMA is better; if based on the four weeks R-square forecast, the T-GARCH is better; if based on the VaR loss function by reality check, the EWMA is better, Finally, The article puts forward some suggestions for future volatility research.