Arrythmia Classification of Electrocardiogram Recorded Data with Random Forest Method

Abstract

Arrythmia is a condition, that our heart beat rhythm change irregularly. The doctor do manual classification process to analyze and diagnose the heart beat rhythm from ECG record. We proposed Random Forest method as classification method, to solve the problem. To cut the preprocessing time, we use WFDB library. INCART arrythmia database is provided as our training and testing data to build the classification model. The feature use is QRS amplitude, QRS amplitude Forward, QRS amplitude Backward, RR interval, Heart Rate Variance (HRV) on the QRS point, backward and forward.We using Scikit Learn for build our classification model, and tested using Scikit Learn and Weka. To classify our object data, we using FOGD-based QRS detector, and to provide our object dataset, Bitalino machine are used. The result still under consideration and need to validate by physician or cardiologist.