New Estimator for an Unknown Mean Gaussian AR(1) Process with Additive Outliers

This paper presents a new estimator for an unknown mean Gaussian AR(1) process with additive outliets. We apply the recursive median adjustment to the weighted symmetric estimator of Park and Fuller [1]. The mean square error (MSE) is derived. Simulation is used to investigate the behavior of this new estimator ((rho) over cap (RMD-uv)) compared to the weighted symmetric estimator ((rho) over cap (uv)) and the recursively mean adjusted weighted symmetric estimator ((rho) over cap (R-uv)) proposed by Niwitpong [2]. Simulation results have shown that the proposed estimator, provides a MSE lower than those of (rho) over cap (uv) and (rho) over cap (R-uv) for almost all situations.