We have built a Heart Disease Detection system using Decision Trees, Random Forest, Gradient Bossting and a Neural Network. We are doing all of this while using fastai for some of the mundane tasks such as data cleaning, feature engineering, data visualization and model evaluation. Decision trees were used to great effect in building our classifiers. We have reached a maximum accuracy of 88% for the decision trees, 90% accuracy for Random Forest and 55.75% with neural nets.
https://medium.com/@nandangrover/heart-disease-detection-using-fastai-and-sklearn-a27df70ed19f