Feature Detection Algorithm Combined with Machine Learning Applied to Abnormal ECG Diagnosis System

With the increasing enthusiasm of machine learning for various human biosignal acquisition and processing, many researchers have devoted themselves to the research of electrocardiogram (ECG) diagnosis incorporating machine learning. In this paper, a lightweight ECG diagnostic system combining Pan-Tompkins (PT) and machine learning algorithms is developed from a wearable ECG monitoring device. The hardware part is small and portable with good endurance; the algorithm part adopts the PT algorithm with 98.63% accuracy for QRS wave group detection. The machine learning (ML) algorithm based on linear regression can also diagnose the loaded ECG signals in the database after the model construction of the training set.