过程控制图在股票收益波动分析中的应用研究

Study on the Application of Statistical Control Chart to Monitoring and Analyzing the Fluctuation of the Series of Stock Proceeds

应用AR型残差控制图解决过程自相关问题;用受控过程的条件标准差替代无条件标准差,构造控制图的控制线,解决过程波动簇聚问题。在实证中,以上证50指数中股票的周收益序列为样本过程,根据其统计性质将其分为四类,并分别举例、选择适当的控制图加以监控。然后,对控制图识别出的异常点举例进行验证、分析,以评价控制图应用于股票收益序列监控的有效性。

In this paper, the AR type residual chart, which combines AR model and residual chart, is proposed to solve the problem of autocorrelation in processes, and the ARCH type control charts, which construct the control chart＇s control lines with the conditional standard variance of sample process rather than with the unconditional standard variance, are proposed to solve the problem of volatility clustering in processes. An empirical study on the application of the control charts to monitoring the series of stock＇s weekly proceeds is carried out. All sample processes are divided into 4 categories according to their statistical characters. For each of the categories an example is cited and a proper control chart is designed. Then some out-of-control signals identified by control charts are testified and analyzed to demonstrate the efficiency of control charts in this empirical study.