2.50
Hdl Handle:
http://hdl.handle.net/10547/293062
Title:
Prediction of survival probabilities with Bayesian Decision Trees
Authors:
Schetinin, Vitaly ( 0000-0003-1826-0153 ) ; Jakaite, Livija; Krzanowski, Wojtek
Abstract:
Practitioners use Trauma and Injury Severity Score (TRISS) models for predicting the survival probability of an injured patient. The accuracy of TRISS predictions is acceptable for patients with up to three typical injuries, but unacceptable for patients with a larger number of injuries or with atypical injuries. Based on a regression model, the TRISS methodology does not provide the predictive density required for accurate assessment of risk. Moreover, the regression model is difficult to interpret. We therefore consider Bayesian inference for estimating the predictive distribution of survival. The inference is based on decision tree models which recursively split data along explanatory variables, and so practitioners can understand these models. We propose the Bayesian method for estimating the predictive density and show that it outperforms the TRISS method in terms of both goodness-of-fit and classification accuracy. The developed method has been made available for evaluation purposes as a stand-alone application.
Affiliation:
University of Bedfordshire; University of Exeter
Citation:
Schetinin, V., Jakaite, L. & Krzanowski, W.J. (2013) 'Prediction of survival probabilities with bayesian decision trees', Expert Systems with Applications, 40 (14), pp.5466-5476.
Publisher:
Elsevier
Journal:
Expert Systems with Applications
Issue Date:
2013
URI:
http://hdl.handle.net/10547/293062
DOI:
10.1016/j.eswa.2013.04.009
Additional Links:
http://www.sciencedirect.com/science/article/pii/S0957417413002467
Type:
Article
Language:
en
Appears in Collections:
Centre for Research in Distributed Technologies (CREDIT)

Full metadata record

DC FieldValue Language
dc.contributor.authorSchetinin, Vitalyen_GB
dc.contributor.authorJakaite, Livijaen_GB
dc.contributor.authorKrzanowski, Wojteken_GB
dc.date.accessioned2013-05-30T13:03:42Z-
dc.date.available2013-05-30T13:03:42Z-
dc.date.issued2013-
dc.identifier.citationSchetinin, V., Jakaite, L. & Krzanowski, W.J. (2013) 'Prediction of survival probabilities with bayesian decision trees', Expert Systems with Applications, 40 (14), pp.5466-5476.en_GB
dc.identifier.doi10.1016/j.eswa.2013.04.009-
dc.identifier.urihttp://hdl.handle.net/10547/293062-
dc.description.abstractPractitioners use Trauma and Injury Severity Score (TRISS) models for predicting the survival probability of an injured patient. The accuracy of TRISS predictions is acceptable for patients with up to three typical injuries, but unacceptable for patients with a larger number of injuries or with atypical injuries. Based on a regression model, the TRISS methodology does not provide the predictive density required for accurate assessment of risk. Moreover, the regression model is difficult to interpret. We therefore consider Bayesian inference for estimating the predictive distribution of survival. The inference is based on decision tree models which recursively split data along explanatory variables, and so practitioners can understand these models. We propose the Bayesian method for estimating the predictive density and show that it outperforms the TRISS method in terms of both goodness-of-fit and classification accuracy. The developed method has been made available for evaluation purposes as a stand-alone application.en_GB
dc.language.isoenen
dc.publisherElsevieren_GB
dc.relation.urlhttp://www.sciencedirect.com/science/article/pii/S0957417413002467-
dc.subjectG760 Machine Learningen_GB
dc.subjectBayesian methodsen_GB
dc.titlePrediction of survival probabilities with Bayesian Decision Treesen
dc.typeArticleen
dc.contributor.departmentUniversity of Bedfordshireen_GB
dc.contributor.departmentUniversity of Exeteren_GB
dc.identifier.journalExpert Systems with Applicationsen_GB
This item is licensed under a Creative Commons License
Creative Commons
All Items in UOBREP are protected by copyright, with all rights reserved, unless otherwise indicated.