Download:
by C. A. Mcgrory, D. M. Titterington
http://www.stats.gla.ac.uk/research/TechRep2006/06-03.pdf
Add To MetaCart
Abstract:
The variational approach to Bayesian inference enables simultaneous estimation of model parameters and model complexity. An interesting feature of this approach is that it appears also to lead to an automatic choice of model complexity. Empirical results from the analysis of hidden Markov models with Gaussian observation densities illustrate this. If the variational algorithm is initialised with a large number of hidden states, redundant states are eliminated as the method converges to a solution, thereby leading to an automatic selection of the number of hidden states. In addition, through the use of a variational approximation, the Deviance Information Criterion (DIC) for Bayesian model selection can be extended to the hidden Markov model framework. Calculation of the DIC provides a further tool for model selection which can be used in conjunction with the variational approach.
Citations
|
2221
|
A tutorial on hidden Markov models and selected applications in speech recognition
– Rabiner
- 1989
|
|
434
|
A maximization technique occuring in the statistical analysis of probabilistic functions of Markov chains
– Baum, Petrie, et al.
- 1970
|
|
426
|
Introduction to variational methods for graphical models
– Jordan, Ghahramani, et al.
- 1999
|
|
415
|
Reversible jump Markov chain Monte Carlo computation and Bayesian model determination
– Green
- 1995
|
|
249
|
On Bayesian analysis of mixtures with an unknown number of components (with discussion
– Richardson, Green
- 1997
|
|
85
|
Inferring parameters and structure of latent variable models by variational Bayes
– Attias
- 1999
|
|
39
|
Bayesian measures of model complexity and fit (with discussion
– Spiegelhalter, Best, et al.
- 2002
|
|
37
|
Ensemble Learning for Hidden Markov Models
– MacKay
- 1997
|
|
31
|
Calculating posterior distributions and modal estimates in Markov mixture models
– Chib
- 1996
|
|
27
|
Bayesian inference in hidden Markov models through the reversible jump Markov chain Monte Carlo method
– Robert, Rydén, et al.
|
|
25
|
Tutorial on variational approximation methods
– Jaakkola
- 2000
|
|
22
|
Variational message passing
– Winn, Bishop
- 2005
|
|
19
|
Variational Bayesian model selection for mixture distributions
– Corduneanu, Bishop
- 2001
|
|
18
|
Hidden Markov models and disease mapping
– Green, Richardson
- 2002
|
|
16
|
Bayesian estimation of hidden Markov chains: A stochastic implementation
– Robert, Celeux, et al.
- 1993
|
|
13
|
Markov chain Monte Carlo methods for computing Bayes factors: A comparative review
– Han, Carlin
- 2001
|
|
12
|
Stylized Facts of Daily Return Series and the Hidden Markov Model
– Ryden, Teräsverta, et al.
- 1998
|
|
9
|
Deviance information criteria for missing data models
– CELEUX, FORBES, et al.
- 2006
|
|
9
|
Hidden Markov Models and Other Models for Discrete-valued Time Series
– MacDonald, Zucchini
- 1997
|
|
8
|
Deviance information criterion for comparing stochastic volatility models
– Berg, Meyer, et al.
- 2004
|
|
5
|
On white noises driven by hidden Markov chains
– Francq, Roussignol
- 1997
|
|
5
|
Local minima, symmetry-breaking, and model pruning in variational free energy minimization. Available online at http://www. inference.phy.cam.ac.uk/mackay/BayesVar.html
– MacKay
- 2001
|
|
5
|
Resampling schemes for hidden Markov models and their application for maximum likelihood estimation
– Robert, Titterington
- 1996
|
|
4
|
Convergence properties of a general algorithm for calculating variational Bayesian estimates for a normal mixture model
– Wang, Titterington
- 2006
|
|
3
|
A Bayesian approach DNA sequence segmentation (with Discussion
– Boys, Henderson
- 2004
|
|
3
|
Variational inference and learning for segmental switching state space models of hidden speech dynamics
– Lee, Attias, et al.
- 2003
|
|
2
|
Variational Approximations in Bayesian Model Selection for Finite Mixture Distributions
– McGrory, Titterington
- 2006
|
