Constraing-Based Learning for Continous-Time Bayesian Networks

Alessandro Bregoli, Marco Scutari, Fabio Stella
Proceedings of the 10th International Conference on Probabilistic Graphical Models, PMLR 138:41-52, 2020.

Abstract

Dynamic Bayesian networks have been well explored in the literature as discrete-time models; however, their continuous-time extensions have seen comparatively little attention. In this paper, we propose the first constraint-based algorithm for learning the structure of continuous-time Bayesian networks. We discuss the different statistical tests and the underlying hypotheses used by our proposal to establish conditional independence. Finally, we validate its performance using synthetic data, and discuss its strengths and limitations. We find that score-based is more accurate in learning networks with binary variables, while our constraint-based approach is more accurate with variables assuming more than two values. However, more experiments are needed for confirmation.

Cite this Paper


BibTeX
@InProceedings{pmlr-v138-bregoli20a, title = {Constraing-Based Learning for Continous-Time Bayesian Networks}, author = {Bregoli, Alessandro and Scutari, Marco and Stella, Fabio}, booktitle = {Proceedings of the 10th International Conference on Probabilistic Graphical Models}, pages = {41--52}, year = {2020}, editor = {Jaeger, Manfred and Nielsen, Thomas Dyhre}, volume = {138}, series = {Proceedings of Machine Learning Research}, month = {23--25 Sep}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v138/bregoli20a/bregoli20a.pdf}, url = {http://proceedings.mlr.press/v138/bregoli20a.html}, abstract = {Dynamic Bayesian networks have been well explored in the literature as discrete-time models; however, their continuous-time extensions have seen comparatively little attention. In this paper, we propose the first constraint-based algorithm for learning the structure of continuous-time Bayesian networks. We discuss the different statistical tests and the underlying hypotheses used by our proposal to establish conditional independence. Finally, we validate its performance using synthetic data, and discuss its strengths and limitations. We find that score-based is more accurate in learning networks with binary variables, while our constraint-based approach is more accurate with variables assuming more than two values. However, more experiments are needed for confirmation.} }
Endnote
%0 Conference Paper %T Constraing-Based Learning for Continous-Time Bayesian Networks %A Alessandro Bregoli %A Marco Scutari %A Fabio Stella %B Proceedings of the 10th International Conference on Probabilistic Graphical Models %C Proceedings of Machine Learning Research %D 2020 %E Manfred Jaeger %E Thomas Dyhre Nielsen %F pmlr-v138-bregoli20a %I PMLR %P 41--52 %U http://proceedings.mlr.press/v138/bregoli20a.html %V 138 %X Dynamic Bayesian networks have been well explored in the literature as discrete-time models; however, their continuous-time extensions have seen comparatively little attention. In this paper, we propose the first constraint-based algorithm for learning the structure of continuous-time Bayesian networks. We discuss the different statistical tests and the underlying hypotheses used by our proposal to establish conditional independence. Finally, we validate its performance using synthetic data, and discuss its strengths and limitations. We find that score-based is more accurate in learning networks with binary variables, while our constraint-based approach is more accurate with variables assuming more than two values. However, more experiments are needed for confirmation.
APA
Bregoli, A., Scutari, M. & Stella, F.. (2020). Constraing-Based Learning for Continous-Time Bayesian Networks. Proceedings of the 10th International Conference on Probabilistic Graphical Models, in Proceedings of Machine Learning Research 138:41-52 Available from http://proceedings.mlr.press/v138/bregoli20a.html.

Related Material