Prediction of probability of survival in critically ill patients optimizing the area under the ROC curve

Authors:
Oscar Luaces;José R. Quevedo;Francisco Taboada;Guillermo M. Albaiceta;Antonio Bahamonde
Affiliations:
Artificial Intelligence Center, University of Oviedo at Gijón, Asturias, Spain;Artificial Intelligence Center, University of Oviedo at Gijón, Asturias, Spain;Hospital Univ. Central de Asturias, University of Oviedo, Asturias, Spain;Hospital Univ. Central de Asturias, University of Oviedo, Asturias, Spain;Artificial Intelligence Center, University of Oviedo at Gijón, Asturias, Spain
Venue:
IJCAI'07 Proceedings of the 20th international joint conference on Artifical intelligence
Year:
2007

Citing 4
Cited 1

Optimizing search engines using clickthrough data

Proceedings of the eighth ACM SIGKDD international conference on Knowledge discovery and data mining
Feature subset selection for learning preferences: a case study

ICML '04 Proceedings of the twenty-first international conference on Machine learning
A support vector method for multivariate performance measures

ICML '05 Proceedings of the 22nd international conference on Machine learning
Active learning for probability estimation using jensen-shannon divergence

ECML'05 Proceedings of the 16th European conference on Machine Learning

Predicting the probability of survival in intensive care unit patients from a small number of variables and training examples

Artificial Intelligence in Medicine

Quantified Score

Hi-index	0.00

Visualization

Abstract

The paper presents a support vector method for estimating probabilities in a real world problem: the prediction of probability of survival in critically ill patients. The standard procedure with Support Vectors Machines uses Platt's method to fit a sigmoid that transforms continuous outputs into probabilities. The method proposed here exploits the difference between maximizing the AUC and minimizing the error rate in binary classification tasks. The conclusion is that it is preferable to optimize the AUC first (using a multivariate SVM) to then fit a sigmoid. We provide experimental evidence in favor of our proposal. For this purpose, we used data collected in general ICUs at 10 hospitals in Spain; 6 of these include coronary patients, while the other 4 do not treat coronary diseases. The total number of patients considered in our study was 2501.