The numbers that matter most are sometimes the ones a person never thinks to check. Heart Disease Prediction Model estimates the likelihood of heart failure from clinical signals like palpitations, resting heart rate, and cholesterol, turning routine measurements into an early warning.
It learns from a dataset of over 10,000 patient records (sourced from an IBM coding challenge) and doesn’t bet everything on a single approach. The project trains and compares several models, from straightforward binary classification to decision trees and random forests, so the strongest predictor earns its place rather than being assumed. A preprocessing module handles the messy work of parsing and cleaning the patient data before any model sees it.
There’s also a server component, a nod toward the part people forget: a model is only useful once something can actually call it.
Built with Python.