A Method for Prediction of Acute Hypotensive Episodes in ICU via PSO and K-Means

Abstract

At present many hospitals have to deal with the patient's care and nursing for Acute Hypotensive Episodes (AHE) in the Intensive Care Unit (ICU). But the prediction of clinical AHE largely depends on the doctor's experience. It is meaningful for clinical care if we can use appropriate methods to predict the AHE in advance and automatically. In this paper, we propose a method to predict the AHE that uses the particle swarm optimization (PSO) algorithm to optimize the initial cluster centers of K-means which extracts the features of patient's mean arterial blood pressure (MAP). Then these features extracted from K-means coupled with the average of a sequence of MAP signal are classified with the support vector machine (SVM). We classified 2863 records, and the best accuracy achieved for the prediction based on the method proposed in this work was 81.2% (sensitivity=83.2% and specificity=80.4%).