摘要
针对高频金融数据收益率序列的厚尾和偏斜性,建立了偏t误差分布假设下的R-GARCH(1,2)模型,对上证380指数5 min频率的高频数据进行了Va R预测,并与经典的正态分布和t分布误差假设下的R-GARCH(1,2)模型的预测精度进行了对比分析。结果表明,误差项服从偏t分布的R-GARCH(1,2)模型能够有效识别上证380指数收益率序列的分布特征,并且能够精确地测量其收益风险。
We built the realized GARCH（ 1,2） model with skewed student＇ s t distribution to forecast VaR of Shanghai Stock Exchange 380 index for the high-frequency financial data with a fat tail and a- symmetry. The model with normal distribution and student＇ s t distribution was used for comparison. The empirical results show that the realized GARCH model with the skewed student＇ s t distribution can identify the characteristics of the return series of Shanghai Stock Exchange 380 index and measure the risk of Shanghai Stock Exchange 380 index more accurately.
出处
《重庆理工大学学报（自然科学）》
CAS
2015年第5期137-141,共5页
Journal of Chongqing University of Technology：Natural Science
基金
重庆市自然科学基金资助项目(cstc2012jj A00018)
重庆市教委科学技术研究项目(KJ130810)
重庆市高等教育教学改革研究项目(1203053)
