Predict sepsis level in intensive medicine : data mining approach

Abstract

This paper aims to support doctor’s decision-making on predicting the Sepsis level. Thus, a set of Data Mining (DM) models were developed using prevision techniques and classification models. These models enable a better doctor’s decision having into account the Sepsis level of the patient. The DM models use real data collected from the Intensive Care Unit of the Santo António Hospital, in Oporto, Portugal. Classification DM models were considered to predict sepsis level in a supervised learning approach. The models were induced making use of the following algorithms: Decision Trees, Support Vector Machines and Naïve Bayes classifier. The models were assessed using the Confusion Matrix, associated metrics, and Cross-validation. The analysis of the total error rate, sensitivity, specificity and accuracy were the metrics used to identify the most relevant measures to predict sepsis level. This work demonstrates that it is possible to predict with great accuracy the sepsis level.