Electrocardiogram (ECG) Circuit Design and Using the Random Forest to ECG Arrhythmia Classification

The paper presents a proposed hardware device for ECG signals measurement with an integrated remote system for signal automatic analysis to help doctors monitor health, diagnose cardiovascular diseases. This measuring device can transmit the ECG signals online to the server, The server is equipped with a software for analyzing ECG signals and classifying them to detect the arrhythmias using a RF (Random Forest) network. The Hermite basis functions were used to generate the feature vectors together with 2 time-based features: the last R-R period and the average of the last 10 R-R periods. The proposed solution was tested with ECG signals taken from databases MIT-BIH (Massachusetts Institute of Technology, Boston’s Beth Israel Hospital). Seven types of ECG beats were classified with an error of 2.28%. The proposed Random Forest algorithm works very fast, which makes it suitable for the requirement of quick classification of some cardiovascular diseases.