Time-Series Bert for Sepsis Detection: Uncovering Patient Trajectories Through Vital Sign Embeddings.

Minassian, R; Radhakrishnan, M; Yu, K

Time-Series Bert for Sepsis Detection: Uncovering Patient Trajectories Through Vital Sign Embeddings.

Minassian, R Radhakrishnan, M

Yu, K

Permalink

Publisher:: IOS Press
Publication Type:: Chapter
Citation:: MEDINFO 2025 — Healthcare Smart × Medicine Deep, 2025, 329, pp. 830-835
Issue Date:: 2025-08-07

Open Access

Copyright Clearance Process

Recently Added
In Progress
Open Access

This item is open access.

Adobe PDF

Download Published versionAdobe PDF (602.03 kB)

View on publisher's site

View statistics

Full metadata record

Field	Value	Language
dc.contributor.author	Minassian, R
dc.contributor.author	Radhakrishnan, M https://orcid.org/0000-0002-9884-1769
dc.contributor.author	Yu, K https://orcid.org/0000-0001-5138-6749
dc.date.accessioned	2025-10-20T05:00:36Z
dc.date.available	2025-10-20T05:00:36Z
dc.date.issued	2025-08-07
dc.identifier.citation	MEDINFO 2025 — Healthcare Smart × Medicine Deep, 2025, 329, pp. 830-835
dc.identifier.uri	http://hdl.handle.net/10453/190526
dc.description.abstract	This study adapts BERT for vital sign time-series analysis in sepsis detection. Using MIMIC-III data, our model's embeddings reveal patient clusters that partition septic from non-septic cases while capturing physiological complexity through diagnosis count distributions. The BERT-based classifier achieves robust performance in both Precision-Recall Area Under Curve (PR AUC), measuring precision maintenance across recall thresholds, and Receiver Operating Characteristic Area Under Curve (ROC AUC), quantifying septic/non-septic case discrimination. Unsupervised learning reveals patient subgroups with distinct physiological profiles, highlighting transformer architectures' ability to extract meaningful patterns from medical time-series for enhanced sepsis monitoring.
dc.format	Print
dc.language	en
dc.publisher	IOS Press
dc.relation	Commonwealth Department of Education
dc.relation	DHCRC-0086
dc.relation	A7930795
dc.relation.ispartof	MEDINFO 2025 — Healthcare Smart × Medicine Deep
dc.relation.ispartofseries	Studies in Health Technology and Informatics
dc.relation.isbasedon	10.3233/SHTI250956
dc.rights	info:eu-repo/semantics/openAccess
dc.subject	0807 Library and Information Studies, 1117 Public Health and Health Services
dc.subject.classification	Medical Informatics
dc.subject.classification	4203 Health services and systems
dc.subject.classification	4601 Applied computing
dc.subject.mesh	Diagnosis, Computer-Assisted
dc.subject.mesh	Vital Signs
dc.subject.mesh	Time Factors
dc.subject.mesh	Algorithms
dc.subject.mesh	Machine Learning
dc.subject.mesh	Sepsis
dc.subject.mesh	Databases, Factual
dc.subject.mesh	ROC Curve
dc.subject.mesh	Critical Care
dc.subject.mesh	Artificial Intelligence
dc.subject.mesh	Humans
dc.subject.mesh	Pattern Recognition, Automated
dc.subject.mesh	Humans
dc.subject.mesh	Sepsis
dc.subject.mesh	Diagnosis, Computer-Assisted
dc.subject.mesh	Critical Care
dc.subject.mesh	ROC Curve
dc.subject.mesh	Algorithms
dc.subject.mesh	Time Factors
dc.subject.mesh	Artificial Intelligence
dc.subject.mesh	Databases, Factual
dc.subject.mesh	Pattern Recognition, Automated
dc.subject.mesh	Vital Signs
dc.subject.mesh	Machine Learning
dc.subject.mesh	Humans
dc.subject.mesh	Sepsis
dc.subject.mesh	Vital Signs
dc.subject.mesh	Diagnosis, Computer-Assisted
dc.subject.mesh	Machine Learning
dc.title	Time-Series Bert for Sepsis Detection: Uncovering Patient Trajectories Through Vital Sign Embeddings.
dc.type	Chapter
utslib.citation.volume	329
utslib.for	0807 Library and Information Studies
utslib.for	1117 Public Health and Health Services
pubs.organisational-group	University of Technology Sydney
pubs.organisational-group	University of Technology Sydney/Faculty of Engineering and Information Technology
pubs.organisational-group	University of Technology Sydney/UTS Groups
pubs.organisational-group	University of Technology Sydney/UTS Groups/Data Science Institute (DSI)
pubs.organisational-group	University of Technology Sydney/UTS Groups/The Trustworthy Digital Society
utslib.copyright.status	open_access	*
dc.rights.license	This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0). To view a copy of this license, visit https://creativecommons.org/licenses/by-nc/4.0/
dc.date.updated	2025-10-20T05:00:34Z
pubs.publication-status	Published
pubs.volume	329

Abstract:

This study adapts BERT for vital sign time-series analysis in sepsis detection. Using MIMIC-III data, our model's embeddings reveal patient clusters that partition septic from non-septic cases while capturing physiological complexity through diagnosis count distributions. The BERT-based classifier achieves robust performance in both Precision-Recall Area Under Curve (PR AUC), measuring precision maintenance across recall thresholds, and Receiver Operating Characteristic Area Under Curve (ROC AUC), quantifying septic/non-septic case discrimination. Unsupervised learning reveals patient subgroups with distinct physiological profiles, highlighting transformer architectures' ability to extract meaningful patterns from medical time-series for enhanced sepsis monitoring.

Please use this identifier to cite or link to this item:

http://hdl.handle.net/10453/190526