Discriminative Switching Linear Dynamical Systems applied to Physiological Condition Monitoring

Konstantinos Georgatzis, Christopher Williams
Proceedings of the 31st Conference on Uncertainty in Artificial Intelligence, PMLR R13:199-208, 2015.

Abstract

We present a Discriminative Switching Linear Dynamical System (DSLDS) applied to patient monitoring in Intensive Care Units (ICUs). Our approach is based on identifying the state-of-health of a patient given their observed vital signs using a discriminative classifier, and then inferring their underlying physiological values conditioned on this status. The work builds on the Factorial Switching Linear Dynamical System (FSLDS) (Quinn et al., 2009) which has been previously used in a similar setting. The FSLDS is a generative model, whereas the DSLDS is a discriminative model. We demonstrate on two real-world datasets that the DSLDS is able to outperform the FSLDS in most cases of interest, and that an alpha-mixture of the two models achieves higher performance than either of the two models separately.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR13-georgatzis15a, title = {Discriminative Switching Linear Dynamical Systems applied to Physiological Condition Monitoring}, author = {Georgatzis, Konstantinos and Williams, Christopher}, booktitle = {Proceedings of the 31st Conference on Uncertainty in Artificial Intelligence}, pages = {199--208}, year = {2015}, editor = {Meila, Marina and Heskes, Tom}, volume = {R13}, series = {Proceedings of Machine Learning Research}, month = {12--16 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/r13/main/assets/georgatzis15a/georgatzis15a.pdf}, url = {https://proceedings.mlr.press/r13/georgatzis15a.html}, abstract = {We present a Discriminative Switching Linear Dynamical System (DSLDS) applied to patient monitoring in Intensive Care Units (ICUs). Our approach is based on identifying the state-of-health of a patient given their observed vital signs using a discriminative classifier, and then inferring their underlying physiological values conditioned on this status. The work builds on the Factorial Switching Linear Dynamical System (FSLDS) (Quinn et al., 2009) which has been previously used in a similar setting. The FSLDS is a generative model, whereas the DSLDS is a discriminative model. We demonstrate on two real-world datasets that the DSLDS is able to outperform the FSLDS in most cases of interest, and that an alpha-mixture of the two models achieves higher performance than either of the two models separately.}, note = {Reissued by PMLR on 04 October 2026.} }
Endnote
%0 Conference Paper %T Discriminative Switching Linear Dynamical Systems applied to Physiological Condition Monitoring %A Konstantinos Georgatzis %A Christopher Williams %B Proceedings of the 31st Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2015 %E Marina Meila %E Tom Heskes %F pmlr-vR13-georgatzis15a %I PMLR %P 199--208 %U https://proceedings.mlr.press/r13/georgatzis15a.html %V R13 %X We present a Discriminative Switching Linear Dynamical System (DSLDS) applied to patient monitoring in Intensive Care Units (ICUs). Our approach is based on identifying the state-of-health of a patient given their observed vital signs using a discriminative classifier, and then inferring their underlying physiological values conditioned on this status. The work builds on the Factorial Switching Linear Dynamical System (FSLDS) (Quinn et al., 2009) which has been previously used in a similar setting. The FSLDS is a generative model, whereas the DSLDS is a discriminative model. We demonstrate on two real-world datasets that the DSLDS is able to outperform the FSLDS in most cases of interest, and that an alpha-mixture of the two models achieves higher performance than either of the two models separately. %Z Reissued by PMLR on 04 October 2026.
APA
Georgatzis, K. & Williams, C.. (2015). Discriminative Switching Linear Dynamical Systems applied to Physiological Condition Monitoring. Proceedings of the 31st Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R13:199-208 Available from https://proceedings.mlr.press/r13/georgatzis15a.html. Reissued by PMLR on 04 October 2026.

Related Material