You are not logged in.

Stabilizing sparse Cox model using clinical structures in electronic medical records

Tran, Truyen, Gopakumar, Shivapratap, Phung, Quoc-Dinh and Venkatesh, Svetha 2014, Stabilizing sparse Cox model using clinical structures in electronic medical records, in IAPR 2014: Proceedings of 2nd International Workshop on Pattern Recognition for Healthcare Analytics, International Association of Pattern Recognition, [Stockholm, Sweden], pp. 1-4.

Attached Files
Name Description MIMEType Size Downloads

Title Stabilizing sparse Cox model using clinical structures in electronic medical records
Author(s) Tran, TruyenORCID iD for Tran, Truyen orcid.org/0000-0001-6531-8907
Gopakumar, Shivapratap
Phung, Quoc-DinhORCID iD for Phung, Quoc-Dinh orcid.org/0000-0002-9977-8247
Venkatesh, Svetha
Conference name Pattern Recognition for Healthcare Analytics. International Workshop (2nd : 2014 : Stockholm, Sweden)
Conference location Stockholm, Sweden
Conference dates 2014/08/24 - 2014/08/24
Title of proceedings IAPR 2014: Proceedings of 2nd International Workshop on Pattern Recognition for Healthcare Analytics
Publication date 2014
Start page 1
End page 4
Total pages 4
Publisher International Association of Pattern Recognition
Place of publication [Stockholm, Sweden]
Summary Stability in clinical prediction models is crucial fortransferability between studies, yet has received littleattention. The problem is paramount in highdimensionaldata which invites sparse models with featureselection capability. We introduce an effectivemethod to stabilize sparse Cox model of time-to-eventsusing clinical structures inherent in Electronic MedicalRecords (EMR). Model estimation is stabilized usinga feature graph derived from two types of EMR structures:temporal structure of disease and intervention recurrences,and hierarchical structure of medical knowledgeand practices. We demonstrate the efficacy of themethod in predicting time-to-readmission of heart failurepatients. On two stability measures – the Jaccardindex and the Consistency index – the use of clinicalstructures significantly increased feature stability withouthurting discriminative power. Our model reporteda competitive AUC of 0.64 (95% CIs: [0.58,0.69]) for 6months prediction.
Language eng
Field of Research 080109 Pattern Recognition and Data Mining
Socio Economic Objective 0 Not Applicable
HERDC Research category E1.1 Full written paper - refereed
Copyright notice ©[2014, The Conference]
Persistent URL http://hdl.handle.net/10536/DRO/DU:30085684

Document type: Conference Paper
Collection: School of Information Technology
Connect to link resolver
 
Unless expressly stated otherwise, the copyright for items in DRO is owned by the author, with all rights reserved.

Versions
Version Filter Type
Citation counts: TR Web of Science Citation Count  Cited 0 times in TR Web of Science
Scopus Citation Count Cited 0 times in Scopus
Google Scholar Search Google Scholar
Access Statistics: 18 Abstract Views, 2 File Downloads  -  Detailed Statistics
Created: Thu, 25 Aug 2016, 17:36:46 EST

Every reasonable effort has been made to ensure that permission has been obtained for items included in DRO. If you believe that your rights have been infringed by this repository, please contact drosupport@deakin.edu.au.