Abstract: The identification of AR processes whose measurement are corrupted by additive noise is considered. A bias compensated least squares (BCLS) algorithm is derived on the framework of solving ...