Minimum uncertainty as Bayesian network model selection principle

Abstract Background Bayesian Network (BN) modeling is a prominent methodology in computational systems biology. However, the incommensurability of datasets frequently encountered in life science domains gives rise to contextual dependence and numerical irregularities in the behavior of model selecti...

Full description

Saved in:
Bibliographic Details
Main Authors: Grigoriy Gogoshin, Andrei S. Rodin
Format: Article
Language:English
Published: BMC 2025-04-01
Series:BMC Bioinformatics
Subjects:
Online Access:https://doi.org/10.1186/s12859-025-06104-5
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849726530249293824
author Grigoriy Gogoshin
Andrei S. Rodin
author_facet Grigoriy Gogoshin
Andrei S. Rodin
author_sort Grigoriy Gogoshin
collection DOAJ
description Abstract Background Bayesian Network (BN) modeling is a prominent methodology in computational systems biology. However, the incommensurability of datasets frequently encountered in life science domains gives rise to contextual dependence and numerical irregularities in the behavior of model selection criteria (such as MDL, Minimum Description Length) used in BN reconstruction. This renders model features, first and foremost dependency strengths, incomparable and difficult to interpret. In this study, we derive and evaluate a model selection principle that addresses these problems. Results The objective of the study is attained by (i) approaching model evaluation as a misspecification problem, (ii) estimating the effect that sampling error has on the satisfiability of conditional independence criterion, as reflected by Mutual Information, and (iii) utilizing this error estimate to penalize uncertainty with the novel Minimum Uncertainty (MU) model selection principle. We validate our findings numerically and demonstrate the performance advantages of the MU criterion. Finally, we illustrate the advantages of the new model evaluation framework on real data examples. Conclusions The new BN model selection principle successfully overcomes performance irregularities observed with MDL, offers a superior average convergence rate in BN reconstruction, and improves the interpretability and universality of resulting BNs, thus enabling direct inter-BN comparisons and evaluations.
format Article
id doaj-art-3db6f1704fc04075bce447b88486f6a9
institution DOAJ
issn 1471-2105
language English
publishDate 2025-04-01
publisher BMC
record_format Article
series BMC Bioinformatics
spelling doaj-art-3db6f1704fc04075bce447b88486f6a92025-08-20T03:10:08ZengBMCBMC Bioinformatics1471-21052025-04-0126112610.1186/s12859-025-06104-5Minimum uncertainty as Bayesian network model selection principleGrigoriy Gogoshin0Andrei S. Rodin1Department of Computational and Quantitative Medicine, Beckman Research Institute, and Diabetes and Metabolism Research Institute, City of Hope National Medical CenterDepartment of Computational and Quantitative Medicine, Beckman Research Institute, and Diabetes and Metabolism Research Institute, City of Hope National Medical CenterAbstract Background Bayesian Network (BN) modeling is a prominent methodology in computational systems biology. However, the incommensurability of datasets frequently encountered in life science domains gives rise to contextual dependence and numerical irregularities in the behavior of model selection criteria (such as MDL, Minimum Description Length) used in BN reconstruction. This renders model features, first and foremost dependency strengths, incomparable and difficult to interpret. In this study, we derive and evaluate a model selection principle that addresses these problems. Results The objective of the study is attained by (i) approaching model evaluation as a misspecification problem, (ii) estimating the effect that sampling error has on the satisfiability of conditional independence criterion, as reflected by Mutual Information, and (iii) utilizing this error estimate to penalize uncertainty with the novel Minimum Uncertainty (MU) model selection principle. We validate our findings numerically and demonstrate the performance advantages of the MU criterion. Finally, we illustrate the advantages of the new model evaluation framework on real data examples. Conclusions The new BN model selection principle successfully overcomes performance irregularities observed with MDL, offers a superior average convergence rate in BN reconstruction, and improves the interpretability and universality of resulting BNs, thus enabling direct inter-BN comparisons and evaluations.https://doi.org/10.1186/s12859-025-06104-5Bayesian networksProbabilistic networksConditional independenceModel selection criteriaMutual informationSampling error
spellingShingle Grigoriy Gogoshin
Andrei S. Rodin
Minimum uncertainty as Bayesian network model selection principle
BMC Bioinformatics
Bayesian networks
Probabilistic networks
Conditional independence
Model selection criteria
Mutual information
Sampling error
title Minimum uncertainty as Bayesian network model selection principle
title_full Minimum uncertainty as Bayesian network model selection principle
title_fullStr Minimum uncertainty as Bayesian network model selection principle
title_full_unstemmed Minimum uncertainty as Bayesian network model selection principle
title_short Minimum uncertainty as Bayesian network model selection principle
title_sort minimum uncertainty as bayesian network model selection principle
topic Bayesian networks
Probabilistic networks
Conditional independence
Model selection criteria
Mutual information
Sampling error
url https://doi.org/10.1186/s12859-025-06104-5
work_keys_str_mv AT grigoriygogoshin minimumuncertaintyasbayesiannetworkmodelselectionprinciple
AT andreisrodin minimumuncertaintyasbayesiannetworkmodelselectionprinciple