School Seminars and Colloquia

Hidden Markov Models with Multiple Observation Processes

ORSUM Seminar

by James Zhao

Institution: The University of Melbourne
Date: Thu 7th August 2008
Time: 1:05 PM
Location: Room 213, Richard Berry Bldg, The University of Melbourne

Abstract: Hidden Markov models traditionally consist of a Markov chain
coupled with a single observation process which is a probabilistic function
of the chain. In this talk, we will consider hidden Markov models with
multiple observation processes, of which only one can be observed at each
point in time. In particular, we will discuss convergence of the information
state under various policies of choosing an observation process, and
optimality of such policies under the criterion of minimisation of long term
expected information entropy of the information state.

For More Information: Mark Fackrell