Heart Disease Prediction

This is a binary classification problem. Several features where used to train a Random Forest model including:

• Age
• RestingBP
• ST Slope

I used Recall as the evaluation metric in this case because I wanted to minimise the number of False Negative cases. That is, patients being told that they are unlikely to have Heart Disease, whilst they are likely to have it. Recall score for the fitted Random Forest model was at 93.2%. Precision was also good at 94.6%

• View the python notebook in GitHub

I used Flask to create a web app for the model and deployed in Heroku. You can use this link to try it out. Just fill in required values and predict: Heart Disease Predictor