BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Early Warning System for Sepsis Detection Using Signature-Based Ma
 chine Learning Models - Hao Ni (University College London\; The Alan Turin
 g Institute)
DTSTART:20210316T103500Z
DTEND:20210316T110000Z
UID:TALK157909@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:Sepsis is a leading cause of death in intensive care\, and ear
 ly detection is needed for timely intervention. However\, methods to ident
 ify the time of sepsis onset from health records vary\, which is critical 
 in retrospective studies. Using the Sepsis-III criteria\, we determine thr
 ee potential onset times for sepsis against which we apply three represent
 ative predictive models: a tree-based model (LGBM)\, a sequential neural n
 etwork model (LSTM)\, and the Cox proportional-hazard model (CoxPHM). Here
  we consider the static demographic factors and the signature feature of p
 hysiological time series of the patients for feature extraction.&nbsp\; Th
 e models were trained on MIMIC-III critical care database. We show that ma
 chine-learning models (LGBM and LSTM) consistently outperformed the classi
 cal approach (CoxPHM). The signature feature set can improve the performan
 ce of the predictive model significantly\, especially the CoxPHM model.<br
 >
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR