摘要
基于蒙特卡洛方法模拟出的股票价格路径分别考察"已实现"核波动(RK)、"已实现"波动(RV)方法的估计精度,结果表明:RK能有效滤出噪音更贴近于真实波动率。进一步将RK与分整自回归移动平均模型结合,并对其分数阶差分算法进行了修改,基于高频数据对我国股票市场的日波动率进行估计和预测。研究结果表明:RK方法在中国市场条件下具有较好的适用性,相对于RV有更好的预测效果。
This paper studies the estimated accuracy of RK and RV methods on the basis of simulated path using Monte Carlo method.Simulation shows that RK method can eliminate the influence of noise effectively and the result of estimation is closer to true volatility.Moreover,the RK method and ARFIMA model are combined to estimate and forecast the volatility in China Stock Markets based on the high-frequency data and modified algorithm of fraction order difference.The results show that RK method has better applicability in China Stock Markets with better predictive validity compared with RV method.
出处
《北京理工大学学报(社会科学版)》
CSSCI
2011年第3期11-15,26,共6页
Journal of Beijing Institute of Technology:Social Sciences Edition
基金
国家自然利学基金资助项目(70771076)
国家杰出青年利学基金资助项目(70225002)
关键词
“已实现”核波动
蒙特卡洛模拟
分整自回归移动平均模型
波动预测
realized kernel
Monte Carlo simulation
autoregressive fractionally integrated moving average model
volatility forecasting