Correntropy based algorithm for mortality prediction in an intensive care unit (ICU)

Abstract

Predicting mortality for ICU patients has become for many years a topic of interest within Health Institutions because ICUs represent large percentage of the budget spent on them. Several scoring systems have been developed to determine the probability of survival of a patient, which allows medical staff making timely decisions on their treatment and the correct distribution of the ICU resources. Alternately to existing scoring systems, this paper proposes a method that had not been used previously for the prediction of mortality, which is the Correntropy.