Bayesian Neural Network in Action in Emergency Care

Miljenko Križmarić & Matej Mertik, University of Maribor , Slovenia,

Abstract.

There are large collections and sets of data about the patients, diagnosis, treatment procedures etc… in the field of medicine today. Using data mining techniques in those cases provide a statistical and logical analysis of the data looking for patterns that can aid by decision making and prediction.
In this paper we presents study of using Bayesian network (BN) in the domain of emergency medicine where BN are especially appropriate because of their
symbolic representation, handling of uncertainity, where different scenarios are possible by given evidences.

We show a use of BN in the case of study out-of-hospital cardiac arrest in emergency medicine, where we use the BN as tool for prediction of return of spontaneous circulation (ROSC) and survival of
hospital discharge.

Based on experiences we conclude the paper withdiscussion of some future applications of using Bayesian networks on human patient simulators in emergency medicine.

Download the full paper Application of Bayesian networks in emergency medicine (25)