金融市场风险测量的混合密度网络模型与应用①
－以香港恒生指数为例
王新宇 魏小平 宋学锋
（中国矿业大学管理学院）
[摘要] 基于混合密度网络模型估计金融时间序列的时变条件密度，提出采用数值模拟方法计算Expected Shortfall的新途径，对香港恒生指数的实证研究表明，VaR的预测精度在高端分位点较好，模型评估指标反映出预测效果良好，混合密度网络可以有效地描述收益的经验分布统计特征和波动规律，是一种有效的金融市场风险测量模型。
关键词 混合密度网络 市场风险测量 后验测试
中图分类号 F830 文献标识码 A
A Model for Measuring Financial Market Risk Based on Mixture Density Networks: a Case of Hong Kong Hangseng Index
Abstract: This paper uses mixture density networks (MDNs) to forecast the time-varying conditional density values of financial time series, and puts forth a new numerical algorithm to calculate Expected Shortfall. The application in Hong Kong Hangseng index approves that Value-at-Risk at high probability levels is accurate, and that it also has good forecasting ability in fluctuation. MDNs can effectively describe the empirical distribution of returns and the volatility mechanism, and is a new model for measuring financial market risk.
Key words: Mixture Density Networks; Market Risk Measure; Back Testing