Analysis of patient outcome using ECG and extreme learning machine ensemble

Abstract

In an acute healthcare setting, the process of assessing severity and assigning appropriate priority of treatment for large numbers of patients is important. Therefore, accurate analysis systems for patient outcome prediction are needed. In this paper, an extreme learning machine (ELM) ensemble based prognosis system is presented for predicting mortality with heart rate variability (HRV) and clinical vital signs. A segment method is implemented to calculate several sets of HRV measures from non-overlapped electrocardiogram segments for each patient and a decision is made through the ELM ensemble.