Parameter Estimation for Generalized Time-varying System with Moving Average Noise

This paper considers the parameter identification of generalized time-varying CAR models with moving average noise. Based on the least squares and the stochastic gradient algorithms are derived to estimate the parameter of generalized time-varying system. The principle of Multi-innovation and a forgetting factor to develop the MI-FFSG algorithm to improve the parameter estimation accuracy. The simulation results verify the effectiveness of the proposed algorithms from the aspects of parameter estimation accuracy and prediction performance.