Tractable Bayesian Inference of Time-Series Dependence Structure

Michael Siracusa, John Fisher III
; Proceedings of the Twelth International Conference on Artificial Intelligence and Statistics, PMLR 5:528-535, 2009.

Abstract

We consider the problem of Bayesian inference over graphical structures describing the interactions among multiple vector time-series. A directed temporal interaction model is presented which assumes a fixed dependence structure among time-series. Using a conjugate prior over this model’s structure and parameters, we focus our attention on characterizing the exact posterior uncertainty in the structure given data. The model is extended via the introduction of a dynamically evolving latent variable which indexes dependence structures over time. Performing inference using this model yields promising results when analyzing the interaction of multiple tracked moving objects.

Cite this Paper


BibTeX
@InProceedings{pmlr-v5-siracusa09a, title = {Tractable Bayesian Inference of Time-Series Dependence Structure}, author = {Michael Siracusa and John Fisher III}, booktitle = {Proceedings of the Twelth International Conference on Artificial Intelligence and Statistics}, pages = {528--535}, year = {2009}, editor = {David van Dyk and Max Welling}, volume = {5}, series = {Proceedings of Machine Learning Research}, address = {Hilton Clearwater Beach Resort, Clearwater Beach, Florida USA}, month = {16--18 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v5/siracusa09a/siracusa09a.pdf}, url = {http://proceedings.mlr.press/v5/siracusa09a.html}, abstract = {We consider the problem of Bayesian inference over graphical structures describing the interactions among multiple vector time-series. A directed temporal interaction model is presented which assumes a fixed dependence structure among time-series. Using a conjugate prior over this model’s structure and parameters, we focus our attention on characterizing the exact posterior uncertainty in the structure given data. The model is extended via the introduction of a dynamically evolving latent variable which indexes dependence structures over time. Performing inference using this model yields promising results when analyzing the interaction of multiple tracked moving objects.} }
Endnote
%0 Conference Paper %T Tractable Bayesian Inference of Time-Series Dependence Structure %A Michael Siracusa %A John Fisher III %B Proceedings of the Twelth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2009 %E David van Dyk %E Max Welling %F pmlr-v5-siracusa09a %I PMLR %J Proceedings of Machine Learning Research %P 528--535 %U http://proceedings.mlr.press %V 5 %W PMLR %X We consider the problem of Bayesian inference over graphical structures describing the interactions among multiple vector time-series. A directed temporal interaction model is presented which assumes a fixed dependence structure among time-series. Using a conjugate prior over this model’s structure and parameters, we focus our attention on characterizing the exact posterior uncertainty in the structure given data. The model is extended via the introduction of a dynamically evolving latent variable which indexes dependence structures over time. Performing inference using this model yields promising results when analyzing the interaction of multiple tracked moving objects.
RIS
TY - CPAPER TI - Tractable Bayesian Inference of Time-Series Dependence Structure AU - Michael Siracusa AU - John Fisher III BT - Proceedings of the Twelth International Conference on Artificial Intelligence and Statistics PY - 2009/04/15 DA - 2009/04/15 ED - David van Dyk ED - Max Welling ID - pmlr-v5-siracusa09a PB - PMLR SP - 528 DP - PMLR EP - 535 L1 - http://proceedings.mlr.press/v5/siracusa09a/siracusa09a.pdf UR - http://proceedings.mlr.press/v5/siracusa09a.html AB - We consider the problem of Bayesian inference over graphical structures describing the interactions among multiple vector time-series. A directed temporal interaction model is presented which assumes a fixed dependence structure among time-series. Using a conjugate prior over this model’s structure and parameters, we focus our attention on characterizing the exact posterior uncertainty in the structure given data. The model is extended via the introduction of a dynamically evolving latent variable which indexes dependence structures over time. Performing inference using this model yields promising results when analyzing the interaction of multiple tracked moving objects. ER -
APA
Siracusa, M. & III, J.F.. (2009). Tractable Bayesian Inference of Time-Series Dependence Structure. Proceedings of the Twelth International Conference on Artificial Intelligence and Statistics, in PMLR 5:528-535

Related Material