Update rules for parameter estimation in Bayesian networks

Update rules for parameter estimation in Bayesian networks (1997)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Eric Bauer , Daphne Koller , Yoram Singer

Citations:

47 - 2 self

Versions

Version History

Metadata Version 3

User correction supplied by SystemCorrections

Datum	Value	Source
TITLE	Update rules for parameter estimation in Bayesian networks	user correction - Legacy Corrections
AUTHOR NAME	Eric Bauer	SVM HeaderParse 0.1
AUTHOR AFFIL	Stanford University	SVM HeaderParse 0.2
AUTHOR NAME	Daphne Koller	SVM HeaderParse 0.1
AUTHOR AFFIL	Stanford University	SVM HeaderParse 0.2
AUTHOR NAME	Yoram Singer	SVM HeaderParse 0.1
AUTHOR AFFIL	AT&T Labs	SVM HeaderParse 0.2
ABSTRACT	This paper re-examines the problem of parameter estimation in Bayesian networks with missing values and hidden variables from the perspective of recent work in on-line learning [12]. We provide a unified framework for parameter estimation that encompasses both on-line learning, where the model is continuously adapted to new data cases as they arrive, and the more traditional batch learning, where a pre-accumulated set of samples is used in a one-time model selection process. In the batch case, our framework encompassesboth the gradient projection algorithm [2, 3] and the EM algorithm [14] for Bayesian networks. The framework also leads to new on-line and batch parameter update schemes, including a parameterized version of EM. We provide both empirical and theoretical results indicating that parameterized EM allows faster convergence to the maximum likelihood parameters than does standard EM. 1 Introduction Over the past few years, there has been a growing interest in the problem of le...	user correction - Legacy Corrections
YEAR	1997	INFERENCE
VENUE TYPE	CONFERENCE	INFERENCE
PAGES	3--13	INFERENCE
CITATIONS	22 found	ParsCit 1.0