Hidden Markov Model Family

From GM-RKB
(Redirected from Hidden Markov Metamodel)
Jump to navigation Jump to search

A Hidden Markov Model Family is a temporal Bayesian metamodel (directed conditional graphical family that is based on a Markov chain) with observable states and hidden states (partial observability).



References

2011

2010

2005

Cannot represent multiple interacting features or long range dependencies between observed elements.

2004

Markov
Models
Do we have control
over the state transitons?
NO YES
Are the states
completely
observable?
YES

Markov Chain

MDP

Markov Decision Process
NO

HMM

Hidden Markov Model

POMDP

Partially Observable
Markov Decision Process

2003

2000

1999

1998

1989

1986