Paper Info
Title
A dynamic Bayesian network approach for time-specific survival probability prediction in patients after ventricular assist device implantation.
Abstract
In this work we present a decision support tool for the calculation of time-dependent survival probability for patients after ventricular assist device implantation. Two different models have been developed, a short term one which predicts survival for the first three months and a long term one that predicts survival for one year after implantation. In order to model the time dependencies between the different time slices of the problem, a dynamic Bayesian network (DBN) approach has been employed. DBNs order to capture the temporal events of the patient disease and the temporal data availability. High accuracy results have been reported for both models. The short and long term DBNs reached an accuracy of 96.97% and 93.55% respectively.
Year
DOI
Venue
2014
10.1109/EMBC.2014.6944296
EMBC
Keywords
Field
DocType
belief networks,decision support tool,time-specific survival probability prediction,diseases,temporal data availability,patient disease,dbn approach,prosthetics,bayes methods,ventricular assist device implantation,dynamic bayesian network approach
Ventricular assist device,Computer science,Artificial intelligence,Survival probability,Machine learning,Dynamic Bayesian network
Conference
Volume
ISSN
Citations 
2014
1557-170X
0
PageRank 
References 
Authors
0.34
7
7
Name
Order
Citations
PageRank
Themis P Exarchos174.66
George Rigas200.34
Y. Goletsis312616.41
Kostas Stefanou400.34
Steven Jacobs5152.44
Maria Giovanna Trivella6468.30
Dimitrios I Fotiadis74914.82