Change Point Estimates for the Parameters of an AR(1) Process

摘要

Three change point estimates for the parameters of a first-order autoregressive (AR(1)) model are developed. The estimators are applied once a control chart signals indicating the presence of a special or assignable cause of variability. The knowledge of the change point can greatly aid the quality practitioner in identifying and removing special causes of variability. Techniques for determining the change point for the location of the AR(1) process, the variance of the white noise process and the autoregressive parameter changed are presented. Simulation results are provided to evaluate the effectiveness of the three change point estimators.