基于“已实现”波动率的ARFIMA模型预测实证研究

摘要

本文采用二次移动平均方法平衡影响“已实现”波动率预测精度的测量误差和市场微观结构误差，利用沪深300指数高频数据实证研究，结果表明“已实现”波动率序列的分布是非正态分布且具有长记忆性，对数“已实现”波动率序列接近于正态分布；最后建ARFIMA模型，并对波动率进行了预测研究。%This paper adopts the method of second moving average to balance measurement error and market microstructure error. The results of empirical study of the behavior of SHSZ300 index show that there is long memory for realized volatility se- ries. The realized volatility series follows non-normal distribution while Logarithm realized volatility follows normal distribution. Finally, ARFIMA model is established to study the distribution characteristics of realized volatility and forecast the future volatility.