A Bayesian Approach to Switching Linear Gaussian State-Space Models for Unsupervised Time-Series Segmentation
Time-series segmentation in the fully unsupervised scenario in which the number of segment-types is a priori unknown is a fundamental problem in many applications. We propose a Bayesian approach to a segmentation model based on the switching linear Gaussian state-space model that enforces a sparse parametrization, such as to use only a small number of a priori available different dynamics to explain the data. This enables us to estimate the number of segment-types within the model, in contrast to previous non-Bayesian approaches where training and comparing several separate models was required. As the resulting model is computationally intractable, we introduce a variational approximation where a reformulation of the problem enables the use of efficient inference algorithms.
| Author(s): | Chiappa, S. |
| Book Title: | ICMLA 2008 |
| Journal: | Proceedings of the 7th International Conference on Machine Learning and Applications (ICMLA 2008) |
| Pages: | 3-9 |
| Year: | 2008 |
| Month: | December |
| Day: | 0 |
| Editors: | Wani, M. A., X.-W. Chen, D. Casasent, L. Kurgan, T. Hu, K. Hafeez |
| Publisher: | IEEE Computer Society |
| Bibtex Type: | Conference Paper (inproceedings) |
| Address: | Los Alamitos, CA, USA |
| DOI: | 10.1109/ICMLA.2008.109 |
| Event Name: | 7th International Conference on Machine Learning and Applications |
| Event Place: | San Diego, CA, USA |
| Digital: | 0 |
| Electronic Archiving: | grant_archive |
| Language: | en |
| Organization: | Max-Planck-Gesellschaft |
| School: | Biologische Kybernetik |
| Links: | |
BibTex
@inproceedings{5380,
title = {A Bayesian Approach to Switching Linear Gaussian State-Space Models for Unsupervised Time-Series Segmentation},
journal = {Proceedings of the 7th International Conference on Machine Learning and Applications (ICMLA 2008)},
booktitle = {ICMLA 2008},
abstract = {Time-series segmentation in the fully unsupervised scenario in which the number of segment-types is a priori unknown is a fundamental problem in many applications. We propose a Bayesian approach to a segmentation model based on the switching linear Gaussian state-space model that enforces a sparse parametrization, such as to use only a small number of a priori available different dynamics to explain the data. This enables us to estimate the number of segment-types within the model, in contrast to previous non-Bayesian approaches where training and comparing several separate models was required. As the
resulting model is computationally intractable, we introduce a variational approximation where a reformulation of the problem enables the use of efficient inference algorithms.},
pages = {3-9},
editors = {Wani, M. A., X.-W. Chen, D. Casasent, L. Kurgan, T. Hu, K. Hafeez},
publisher = {IEEE Computer Society},
organization = {Max-Planck-Gesellschaft},
school = {Biologische Kybernetik},
address = {Los Alamitos, CA, USA},
month = dec,
year = {2008},
slug = {5380},
author = {Chiappa, S.},
month_numeric = {12}
}