AN OPTIMAL FILTERING ALGORITHM FOR PARAMETER ESTIMATION OF BILINEAR TIME SERIES MODEL
-
Graphical Abstract
-
Abstract
A parameter estimation algorithm for bilinear time series model which is based on Kalman filtering is proposed in this paper. To overcome the divergence of the filter, the prediction error covariance matrix is monitered, and is reset when necessary. A simulation example is given to compare this algorithm with the recursive prediction error psrameter estimator.
-
-