自回归模型参数变点的修正残差CUSUM监测

Corrected Residual CUSUM Monitoring of Autoregressive Model Parameter Change Points

基于自回归模型参数变点的在线监测问题,为缩短监视变点的平均运行长度,引进窗宽参数调整监测起始时刻,提高检验势.用最小二乘法构造参数变点的残差累积和监测统计量,得出原假设和备择假设条件下监测统计量的极限性质.

Based on the on-line monitoring problem of the parameter change point of the autoregressive model,in order to shorten the average running length of the monitoring change point,the window width parameter is introduced to adjust the starting time of monitoring and improve the test potential.The residual accumulation and monitoring statistics of the parameter change point are constructed by least square method,and the limit properties of monitoring statistics under hypothesis and alternative hypothesis are obtained.