Heart Arrhythmia Abnormality Classification Using Machine Learning

Heart arrhythmia refers to the medical condition in which the electrical impulses that control the heartbeat become abnormal and ends up to an irregular rhythm. The heart abnormality has been always identified using electrocardiogram which is a signal that represents the heart electrical activities. In this work, we collected different ECG databases from different resources, preprocessed them, and extracted some useful features from the ECG signal. Then, feature selection methods were applied to the dataset that we already extracted, and we ended up with three input attributes and four corresponding outputs normal rhythm, tachycardia disease, bradycardia, and left bundle branch. We fed up our pre-cleaned feature vector into the classifiers for training and testing using different methods SVM, Naïve Bayes, and Random Forest. In the last step, we evaluated our models and compared the result.