Title :
Structured risk-sensitivity for partially observed Markov chains
Author :
Ramezani, Vahid R. ; Marcus, Steven I. ; Fu, Michael
Author_Institution :
Dept. of Electr. & Comput. Eng., Maryland Univ., College Park, MD, USA
Abstract :
Product estimators for partially observed Markov chains are introduced and a notion of risk-sensitivity for which the risk is non-uniform and state dependent is defined. Product probability is introduced and studied in the context of left-to-right Markov chains for uniform and state-dependent cases. We show that the qualitative behavior of these estimators is related to certain threshold properties.
Keywords :
Markov processes; probability; risk analysis; stochastic systems; left-to-right Markov chains; partially observed Markov chains; product estimators; product probability; risk-sensitivity; structured risk-sensitivity; uniform state dependent cases; Contracts; Cost function; Educational institutions; Error correction; Estimation error; Filters; Smoothing methods; State estimation; Stochastic processes; Yttrium;
Conference_Titel :
Decision and Control, 2004. CDC. 43rd IEEE Conference on
Print_ISBN :
0-7803-8682-5
DOI :
10.1109/CDC.2004.1429247
