Learning Bayesian Belief Networks Based on the Minimum Description Length Principle: An Efficient Algorithm Using the B & B Technique Export

by: Joe Suzuki

In International Conference on Machine Learning (1996), pp. 462-470.

[Posts]

View FullText article

CiteSeerX Beta, Rent at DeepDyve

mkuro's tags for this article

bayesian learning network

Reviews [Write a review of this article]

Notes for this article

mkuro has 1 private note and 0 public notes for this article. If you are mkuro then you can log in to see the private note.

Find related articles from these CiteULike users

jrw, mkuro, roys

Find related articles with these CiteULike tags

algorithms, bayesian, bayesian_networks, combinatorics, computation, graphs, learning, mdl, network, score_function

Abstract

This paper addresses the problem of learning Bayesian belief networks (BBN) based on the minimum description length (MDL) principle. First, we give a formula of description length based on which the MDL-based procedure learns a BBN. Secondly, we point out that the difference between the MDL-based and Cooper and Herskovits procedures is essentially in the priors rather than in the approaches (MDL and Bayesian), and recommend a class of priors from which the formula is obtained. Finally, we show...

@inproceedings{citeulike:586537, abstract = {This paper addresses the problem of learning Bayesian belief networks (BBN) based on the minimum description length (MDL) principle. First, we give a formula of description length based on which the MDL-based procedure learns a BBN. Secondly, we point out that the difference between the MDL-based and Cooper and Herskovits procedures is essentially in the priors rather than in the approaches (MDL and Bayesian), and recommend a class of priors from which the formula is obtained. Finally, we show...}, author = {Suzuki, Joe}, booktitle = {International Conference on Machine Learning}, citeulike-article-id = {586537}, citeulike-linkout-0 = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.46.2318}, keywords = {bayesian, learning, network}, pages = {462--470}, posted-at = {2006-04-14 18:19:08}, priority = {3}, title = {Learning Bayesian Belief Networks Based on the Minimum Description Length Principle: An Efficient Algorithm Using the B \& B Technique}, url = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.46.2318}, year = {1996} }

RIS record

TY - CONF ID - citeulike:586537 L3 - citeulike-article-id:586537 N2 - This paper addresses the problem of learning Bayesian belief networks (BBN) based on the minimum description length (MDL) principle. First, we give a formula of description length based on which the MDL-based procedure learns a BBN. Secondly, we point out that the difference between the MDL-based and Cooper and Herskovits procedures is essentially in the priors rather than in the approaches (MDL and Bayesian), and recommend a class of priors from which the formula is obtained. Finally, we show... EP - 470 TI - Learning Bayesian Belief Networks Based on the Minimum Description Length Principle: An Efficient Algorithm Using the B & B Technique T2 - International Conference on Machine Learning SP - 462 KW - bayesian KW - learning KW - network AU - Suzuki, Joe PY - 1996/// UR - http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.46.2318 ER -

bayesian	6
network	5
learning	5
collaborative	2
filtering	2
criterion	1
information	1

Learning Bayesian Belief Networks Based on the Minimum Description Length Principle: An Efficient Algorithm Using the B & B Technique Export

Citation Format