Title :
A randomized EM-algorithm for estimating quantized linear Gaussian regression
Author :
Finesso, Lorenzo ; Gerencsér, László ; Kmecs, Ildikó
Author_Institution :
LADSEB, CNR, Padova, Italy
Abstract :
The purpose of this paper is to formulate and study the problem of system identification with Gaussian noise and quantized observations. Two examples are considered in more details: Gaussian AR(1)-systems and the simplest Gaussian linear regression. The main result of the paper is the development of a randomized version of the EM-method for the effective solution of the likelihood equation. Computational experiments are also given
Keywords :
Gaussian noise; approximation theory; estimation theory; hidden Markov models; identification; Gaussian noise; Metropolis method; hidden Markov models; linear Gaussian regression; quantized observations; randomized EM-algorithm; stochastic approximation; system identification; Automation; Biomedical computing; Biomedical engineering; Convergence; Gaussian noise; Hidden Markov models; Integral equations; Linear regression; Random variables; System identification;
Conference_Titel :
Decision and Control, 1999. Proceedings of the 38th IEEE Conference on
Conference_Location :
Phoenix, AZ
Print_ISBN :
0-7803-5250-5
DOI :
10.1109/CDC.1999.833359
