Enhanced Machine Learning Feature Selection Algorithm for Cardiac Arrhythmia in a Personal Healthcare Application

In this paper, we propose a personal healthcare system to diagnosis, monitor, identify ectopic heartbeats and provide services for heart diseases patients. The proposed system includes three subsystems, (1) the Electrocardiograph signal acquisition device, (2) the mobile device with APP software for signal processing and communication, and (3) Web service for feature extraction, selection and heartbeat classification based on the support vector machine. The proposed system can identify 16 kinds of heartbeats and it can be used in daily life.