基于BSP-HAR-RV模型的最优抽样频率研究

金融高频时间序列由于数量大,周期短,信息丰富从而可以很好的反映金融市场特征。通过绘出平均双幂变差已实现波动率散点图(BiPowe Realized Volatility Signature Plot,BSP),建立BSP-HAR-RV模型,改进以往国内通过列举法选择最优频率的方法。最后对TCL集团股票价格一年多的高频数据实证分析,验证模型结果,并将其在最优频率下测量得到的HAR-RV模型预测结果与以往广泛使用的5min、10min频率下得到的结果进行比较,发现最优抽样频率下模型预测能力较好,具有可行性。

An Optimal Sampling Frequency Studies Based on BSP-HAR-RV Model

High-frequency time series can reflect the characteristics of financial markets more comprehensively because of large quantities, short cycles and abundant information. By drawing an average BiPower realized volatility Signature Plot(BSP) and seting up BSP - HAR - RV model, improve the optimal-frequency-selecting enumerating-method used in the domestic.Then do the empirical analysis to verify the model using the TCL stock price high-frequency data ,and find the model predict ability based on the optimal sampling frequency is good and feasible through the comparison between the BSP-HAR-RV model predicted results based on the optimal sampling frequency and the HAR-RV predicted results based on 5-min- frequencie and 10-min- frequencie.