LMS-LIKE ESTIMATION FOR TIME VARYING PARAMETERS

An LMS-like algorithm is applied for estimating the time-varying parameter θin the linear model y=φ~τθ+v, which is general in the sense that none of the probabilistic properties such as stationarity, Marker property, independence and ergodicity is imposed on any of the processes {y}, {φ}, {θ} and {v}. It is shown that the α-th moment of the estimation error is of order of the α-th moment of the observation noise and the parameter variation