There are many publicly available Machine Learning projects for classifying ECG data. The MIT-BIH Arrhythmia Database is a popular choice to train such models.
Since ECG data can be treated as image data, many of the models are built around a Convolutional Neural Network (CNN). Quite often, Keras is the preferred API.
Inspired by the idea that Tensorflow Estimators remove a lot of the complexities of building out Neural Networks, and the fact that a DNNClassifier works quite well in building image classification models, I decided to try it out on an ECG Classification model, which is available here.
