An improved structural EM to learn dynamic Bayesian nets

Cassio P. de Campos; Zhi Zeng; Qiang Ji

doi:10.1109/ICPR.2010.152

An improved structural EM to learn dynamic Bayesian nets

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

Abstract

This paper addresses the problem of learning structure of Bayesian and Dynamic Bayesian networks from incomplete data based on the Bayesian Information Criterion. We describe a procedure to map the problem of the dynamic case into a corresponding augmented Bayesian network through the use of structural constraints. Because the algorithm is exact and anytime, it is well suitable for a structural Expectation-Maximization (EM) method where the only source of approximation is due to the EM itself. We show empirically that the use a global maximizer inside the structural EM is computationally feasible and leads to more accurate models.

Original language	English
Title of host publication	Proceedings - 2010 20th International Conference on Pattern Recognition, ICPR 2010
Place of Publication	Piscataway
Publisher	Institute of Electrical and Electronics Engineers
Pages	601-604
Number of pages	4
ISBN (Electronic)	978-1-4244-7541-4
ISBN (Print)	978-1-4244-7542-1
DOIs	https://doi.org/10.1109/ICPR.2010.152
Publication status	Published - 18 Nov 2010
Externally published	Yes
Event	2010 20th International Conference on Pattern Recognition, ICPR 2010 - Istanbul, Turkey Duration: 23 Aug 2010 → 26 Aug 2010

Conference

Conference	2010 20th International Conference on Pattern Recognition, ICPR 2010
Country/Territory	Turkey
City	Istanbul
Period	23/08/10 → 26/08/10

Access to Document

10.1109/ICPR.2010.152

Cite this

@inproceedings{330ea7421cda4418b87f3c24d7652763,

title = "An improved structural EM to learn dynamic Bayesian nets",

abstract = "This paper addresses the problem of learning structure of Bayesian and Dynamic Bayesian networks from incomplete data based on the Bayesian Information Criterion. We describe a procedure to map the problem of the dynamic case into a corresponding augmented Bayesian network through the use of structural constraints. Because the algorithm is exact and anytime, it is well suitable for a structural Expectation-Maximization (EM) method where the only source of approximation is due to the EM itself. We show empirically that the use a global maximizer inside the structural EM is computationally feasible and leads to more accurate models.",

author = "{de Campos}, {Cassio P.} and Zhi Zeng and Qiang Ji",

year = "2010",

month = nov,

day = "18",

doi = "10.1109/ICPR.2010.152",

language = "English",

isbn = "978-1-4244-7542-1",

pages = "601--604",

booktitle = "Proceedings - 2010 20th International Conference on Pattern Recognition, ICPR 2010",

publisher = "Institute of Electrical and Electronics Engineers",

address = "United States",

note = "2010 20th International Conference on Pattern Recognition, ICPR 2010 ; Conference date: 23-08-2010 Through 26-08-2010",

}

de Campos, CP, Zeng, Z & Ji, Q 2010, An improved structural EM to learn dynamic Bayesian nets. in Proceedings - 2010 20th International Conference on Pattern Recognition, ICPR 2010., 5597455, Institute of Electrical and Electronics Engineers, Piscataway, pp. 601-604, 2010 20th International Conference on Pattern Recognition, ICPR 2010, Istanbul, Turkey, 23/08/10. https://doi.org/10.1109/ICPR.2010.152

An improved structural EM to learn dynamic Bayesian nets. / de Campos, Cassio P.; Zeng, Zhi; Ji, Qiang.
Proceedings - 2010 20th International Conference on Pattern Recognition, ICPR 2010. Piscataway: Institute of Electrical and Electronics Engineers, 2010. p. 601-604 5597455.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - An improved structural EM to learn dynamic Bayesian nets

AU - de Campos, Cassio P.

AU - Zeng, Zhi

AU - Ji, Qiang

PY - 2010/11/18

Y1 - 2010/11/18

N2 - This paper addresses the problem of learning structure of Bayesian and Dynamic Bayesian networks from incomplete data based on the Bayesian Information Criterion. We describe a procedure to map the problem of the dynamic case into a corresponding augmented Bayesian network through the use of structural constraints. Because the algorithm is exact and anytime, it is well suitable for a structural Expectation-Maximization (EM) method where the only source of approximation is due to the EM itself. We show empirically that the use a global maximizer inside the structural EM is computationally feasible and leads to more accurate models.

AB - This paper addresses the problem of learning structure of Bayesian and Dynamic Bayesian networks from incomplete data based on the Bayesian Information Criterion. We describe a procedure to map the problem of the dynamic case into a corresponding augmented Bayesian network through the use of structural constraints. Because the algorithm is exact and anytime, it is well suitable for a structural Expectation-Maximization (EM) method where the only source of approximation is due to the EM itself. We show empirically that the use a global maximizer inside the structural EM is computationally feasible and leads to more accurate models.

UR - http://www.scopus.com/inward/record.url?scp=78149474100&partnerID=8YFLogxK

U2 - 10.1109/ICPR.2010.152

DO - 10.1109/ICPR.2010.152

M3 - Conference contribution

SN - 978-1-4244-7542-1

SP - 601

EP - 604

BT - Proceedings - 2010 20th International Conference on Pattern Recognition, ICPR 2010

PB - Institute of Electrical and Electronics Engineers

CY - Piscataway

T2 - 2010 20th International Conference on Pattern Recognition, ICPR 2010

Y2 - 23 August 2010 through 26 August 2010

ER -

An improved structural EM to learn dynamic Bayesian nets

Abstract

Conference

Access to Document

Other files and links

Fingerprint

Cite this