Using Applied Mathematics to Identify Electrocardiogram Features

Abstract

This work presents a framework for identifying features on a beat-by-beat basis in electrocardiogram (ECG) signals. Since each feature corresponds to a different part of the cardiac cycle, tracking changes in these features over time can provide insight regarding a patient's clinical status. Using tools from numerical linear algebra to first identify a representative subset of beats from a larger data set, we can then use clinical expertise and data science methods to identify individual beat features.