Learning bayesian network structure for risky behavior modelling

Alena Suvorova, Alexander Tulupyev

Research output

Abstract

Bayesian Belief Networks (BBN) provide a comprehensible framework for representing complex systems that allows including expert knowledge and statistical data simultaneously. We explored BBN models for estimating risky behavior rate and compared several network structures, both expert-based and data-based. To learn and evaluate models we used generated behavior data with 9393 observations. We applied both score-based and constraint-based structure learning algorithms. The score-based structures represented better quality scores according to BIC and log-likelihood, prediction quality was almost the same for data-based models and lower but sufficient for expert-based models. Hence, in case of limited data we can reduce computations and apply expert-based structure for solving practical issues.

Original languageEnglish
Title of host publicationProceedings of the 3rd International Scientific Conference “Intelligent Information Technologies for Industry” (IITI’18) - Volume 2
EditorsValery Tarassov, Sergey Kovalev, Andrey Sukhanov, Ajith Abraham, Vaclav Snasel
PublisherSpringer
Pages58-65
Number of pages8
ISBN (Print)9783030018207
DOIs
Publication statusPublished - 1 Jan 2019
Event3rd International Scientific Conference on Intelligent Information Technologies for Industry, IITI 2018 - Sochi
Duration: 17 Sep 201821 Sep 2018

Publication series

NameAdvances in Intelligent Systems and Computing
Volume875
ISSN (Print)2194-5357

Conference

Conference3rd International Scientific Conference on Intelligent Information Technologies for Industry, IITI 2018
CountryRussian Federation
CitySochi
Period17/09/1821/09/18

Fingerprint

Bayesian networks
Learning algorithms
Large scale systems

Scopus subject areas

  • Control and Systems Engineering
  • Computer Science(all)

Cite this

Suvorova, A., & Tulupyev, A. (2019). Learning bayesian network structure for risky behavior modelling. In V. Tarassov, S. Kovalev, A. Sukhanov, A. Abraham, & V. Snasel (Eds.), Proceedings of the 3rd International Scientific Conference “Intelligent Information Technologies for Industry” (IITI’18) - Volume 2 (pp. 58-65). (Advances in Intelligent Systems and Computing; Vol. 875). Springer. https://doi.org/10.1007/978-3-030-01821-4_7
Suvorova, Alena ; Tulupyev, Alexander. / Learning bayesian network structure for risky behavior modelling. Proceedings of the 3rd International Scientific Conference “Intelligent Information Technologies for Industry” (IITI’18) - Volume 2. editor / Valery Tarassov ; Sergey Kovalev ; Andrey Sukhanov ; Ajith Abraham ; Vaclav Snasel. Springer, 2019. pp. 58-65 (Advances in Intelligent Systems and Computing).
@inproceedings{bdd6c3105dc14e4094913ab1a3ec0c5b,
title = "Learning bayesian network structure for risky behavior modelling",
abstract = "Bayesian Belief Networks (BBN) provide a comprehensible framework for representing complex systems that allows including expert knowledge and statistical data simultaneously. We explored BBN models for estimating risky behavior rate and compared several network structures, both expert-based and data-based. To learn and evaluate models we used generated behavior data with 9393 observations. We applied both score-based and constraint-based structure learning algorithms. The score-based structures represented better quality scores according to BIC and log-likelihood, prediction quality was almost the same for data-based models and lower but sufficient for expert-based models. Hence, in case of limited data we can reduce computations and apply expert-based structure for solving practical issues.",
keywords = "Bayesian belief network, Behavior models, Machine learning, Risky behavior, Structure learning",
author = "Alena Suvorova and Alexander Tulupyev",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/978-3-030-01821-4_7",
language = "English",
isbn = "9783030018207",
series = "Advances in Intelligent Systems and Computing",
publisher = "Springer",
pages = "58--65",
editor = "Valery Tarassov and Sergey Kovalev and Andrey Sukhanov and Ajith Abraham and Vaclav Snasel",
booktitle = "Proceedings of the 3rd International Scientific Conference “Intelligent Information Technologies for Industry” (IITI’18) - Volume 2",
address = "Germany",

}

Suvorova, A & Tulupyev, A 2019, Learning bayesian network structure for risky behavior modelling. in V Tarassov, S Kovalev, A Sukhanov, A Abraham & V Snasel (eds), Proceedings of the 3rd International Scientific Conference “Intelligent Information Technologies for Industry” (IITI’18) - Volume 2. Advances in Intelligent Systems and Computing, vol. 875, Springer, pp. 58-65, Sochi, 17/09/18. https://doi.org/10.1007/978-3-030-01821-4_7

Learning bayesian network structure for risky behavior modelling. / Suvorova, Alena; Tulupyev, Alexander.

Proceedings of the 3rd International Scientific Conference “Intelligent Information Technologies for Industry” (IITI’18) - Volume 2. ed. / Valery Tarassov; Sergey Kovalev; Andrey Sukhanov; Ajith Abraham; Vaclav Snasel. Springer, 2019. p. 58-65 (Advances in Intelligent Systems and Computing; Vol. 875).

Research output

TY - GEN

T1 - Learning bayesian network structure for risky behavior modelling

AU - Suvorova, Alena

AU - Tulupyev, Alexander

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Bayesian Belief Networks (BBN) provide a comprehensible framework for representing complex systems that allows including expert knowledge and statistical data simultaneously. We explored BBN models for estimating risky behavior rate and compared several network structures, both expert-based and data-based. To learn and evaluate models we used generated behavior data with 9393 observations. We applied both score-based and constraint-based structure learning algorithms. The score-based structures represented better quality scores according to BIC and log-likelihood, prediction quality was almost the same for data-based models and lower but sufficient for expert-based models. Hence, in case of limited data we can reduce computations and apply expert-based structure for solving practical issues.

AB - Bayesian Belief Networks (BBN) provide a comprehensible framework for representing complex systems that allows including expert knowledge and statistical data simultaneously. We explored BBN models for estimating risky behavior rate and compared several network structures, both expert-based and data-based. To learn and evaluate models we used generated behavior data with 9393 observations. We applied both score-based and constraint-based structure learning algorithms. The score-based structures represented better quality scores according to BIC and log-likelihood, prediction quality was almost the same for data-based models and lower but sufficient for expert-based models. Hence, in case of limited data we can reduce computations and apply expert-based structure for solving practical issues.

KW - Bayesian belief network

KW - Behavior models

KW - Machine learning

KW - Risky behavior

KW - Structure learning

UR - http://www.scopus.com/inward/record.url?scp=85058550255&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-01821-4_7

DO - 10.1007/978-3-030-01821-4_7

M3 - Conference contribution

AN - SCOPUS:85058550255

SN - 9783030018207

T3 - Advances in Intelligent Systems and Computing

SP - 58

EP - 65

BT - Proceedings of the 3rd International Scientific Conference “Intelligent Information Technologies for Industry” (IITI’18) - Volume 2

A2 - Tarassov, Valery

A2 - Kovalev, Sergey

A2 - Sukhanov, Andrey

A2 - Abraham, Ajith

A2 - Snasel, Vaclav

PB - Springer

ER -

Suvorova A, Tulupyev A. Learning bayesian network structure for risky behavior modelling. In Tarassov V, Kovalev S, Sukhanov A, Abraham A, Snasel V, editors, Proceedings of the 3rd International Scientific Conference “Intelligent Information Technologies for Industry” (IITI’18) - Volume 2. Springer. 2019. p. 58-65. (Advances in Intelligent Systems and Computing). https://doi.org/10.1007/978-3-030-01821-4_7