Learning and selection of dynamic Bayesian networks for online non-stationary process

Apprentissage et sélection de réseaux bayésiens dynamiques pour les processus online non stationnaires

Matthieu Hourbracq (matthieu.hourbracq@gmail.com), Pierre-Henri Wuillemin (pierre-henri.wuillemin@lip6.fr), Christophe Gonzales (christophe.gonzales@lis-lab.fr) and Philippe Baumard (philippe.baumard@lecnam.net)
Additional contact information
Matthieu Hourbracq: Akheros, DECISION - LIP6 - SU - Sorbonne Université - CNRS - Centre National de la Recherche Scientifique
Pierre-Henri Wuillemin: DECISION - LIP6 - SU - Sorbonne Université - CNRS - Centre National de la Recherche Scientifique
Christophe Gonzales: DECISION - LIP6 - SU - Sorbonne Université - CNRS - Centre National de la Recherche Scientifique
Philippe Baumard: LIRSA - Laboratoire interdisciplinaire de recherche en sciences de l'action - Cnam - Conservatoire National des Arts et Métiers [Cnam], ESD R3C - Équipe Sécurité & Défense - Renseignement, Criminologie, Crises, Cybermenaces - Cnam - Conservatoire National des Arts et Métiers [Cnam], Cnam - Conservatoire National des Arts et Métiers [Cnam], Akheros

Post-Print from HAL

Abstract: Dynamic Bayesian Networks (DBNs) provide a principled scheme for modeling and learning conditional dependencies from complex multivariate time-series data. However, in most cases, the underlying generative Markov model is assumed to be homogeneous, mea- ning that neither its topology nor its parameters evolve over time. Therefore, learning a DBN to model a non-stationary process under this assumption will amount to poor predictions capa- bilities. Thus we build a framework to identify, in a streamed manner, transition times between underlying models and a framework to learn them in real time, without assumptions about their evolution. We propose a model for the dynamic of the transitions between modes stemming from Hidden semi-Markov Models (HsMMs) and Graphical Duration Models (GDMs). We show the method performances on simulated datasets.

Keywords: apprentissage; non-stationnaire; tv-DBN; ns-DBN; DBN; temps réel (search for similar items in EconPapers)
Date: 2018
Note: View the original document on HAL open archive server: https://cnam.hal.science/hal-03228681v1
References: View complete reference list from CitEc
Citations:

Published in Revue des Sciences et Technologies de l'Information - Série RIA : Revue d'Intelligence Artificielle, 2018, Réseaux bayésiens et modèles probabilistes, 32 (1), pp. 75-109. ⟨10.3166/RIA.32.75-109⟩

Downloads: (external link)
https://cnam.hal.science/hal-03228681v1/document (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:hal:journl:hal-03228681

DOI: 10.3166/RIA.32.75-109

Access Statistics for this paper

More papers in Post-Print from HAL
Bibliographic data for series maintained by CCSD (hal@ccsd.cnrs.fr).