Project 3 For The Metis Data Science Program
A website that takes diagnostic information from someone experiencing chest pain and shows a gauge of their predicted risk of having heart disease.
The demo version (index.html) that you can run on localhost is a logisitic regression classifier. We also built a voting classifier that uses GaussianNB, SVM, and Logistic regression for the base model. Addional feature models--ones with blood pressure & cholesterol--use GuassianNB and SVM. Predictions are made using a sklearn classifier trained on the Cleveland, OH subset of the UCI Heart Disease Data Set. The model has been tuned to maximize recall (minimize false negatives).
Completed as project 'Mcnulty' for the 2016 Metis San Francisco cohort.
- This is a prototype for demonstration purposes only. Nobody involved in its creation has any medical training whatsoever. If you are experiencing chest pain, consult an actual doctor. *