Title :
Improved smoother dynamics for discrete time HMM parameter estimation
Author :
Elliott, R.J. ; Malcolm, W.P.
Author_Institution :
Fac. of Manage., Calgary Univ., Alta., Canada
Abstract :
In this article, we consider hidden Markov model (HMM) parameter estimation in the context of an expectation maximisation (EM) algorithm. The models we study are discrete-time Markov chains observed in Gaussian noise. New formulae for updating smoothed estimates are given for the models just described. These formulae are computed by exploiting a duality between a forward-in-time unnormalised probability process and its dual
Keywords :
Gaussian noise; discrete time filters; duality (mathematics); hidden Markov models; optimisation; parameter estimation; probability; smoothing methods; Gaussian noise; discrete-time HMM parameter estimation; discrete-time Markov chains; duality; expectation maximisation algorithm; forward-in-time unnormalised probability process; hidden Markov model; martingales; reference probability; smoothed estimate updating; smoother dynamics; Books; Boundary conditions; Equations; Gaussian noise; Glands; Hidden Markov models; Mathematics; Noise robustness; Parameter estimation; Random variables;
Conference_Titel :
Decision and Control, 2001. Proceedings of the 40th IEEE Conference on
Conference_Location :
Orlando, FL
Print_ISBN :
0-7803-7061-9
DOI :
10.1109/.2001.980402
