基于LMSV模型的半参数方法对股市波动长记忆特征的识别

The Long Memory Property Detection on Stock Market Volatility by Semiparametric Methods with LMSV Model

在金融时间序列波动具有显著的长记忆性这一背景之下,研究了LMSV模型长记忆参数的估计问题。首先,分析了LMSV模型的相关性质;接着,根据LMSV模型和ARFIMA模型的良好对应关系,提出了估计LMSV模型长记忆参数的半参数方法;最后,基于股市数据,验证了波动半参数方法的有效性。

Under the background of financial time series with obviously long memory property,estimation the long memory parameter of LMSV model is studied.Firstly,the properties of LMSV model are analyzed.Then the semiparametric methods of estimating long memory parametric in LMSV model are proposed based on the suitable corresponding relationships.Finally,the efficiency of the semiparametric methods is testified by stock market data.