ECG anomaly detection using an interpretable Autoencoder model

When it comes to your heart, getting a second opinion is never a bad thing. We all need assurances and the more people tell us, the better we feel about our decision being right. In medicine, the judgment of doctors over some symptom is either based on a machine or on their experience. A patient can communicate with the doctor to understand how they reached a particular judgement. This is where the machines lack when it comes to providing accuracy and more importantly, transparency. The lack of accurate and interpretable machine models for disease/anomaly detection is a pressing issue and needs to be tackled effectively. This paper aims to target anomaly detection in ECGs through the means of an accurate and interpretable machine learning model.