Heart Disease Prediction

Rising health concerns also demand quick answers to curiosity. To feed your health hunger, this ML web application uses RandomForestClassifier to give 87% accuracy in prediction integrating sqlite3.

Comment 1

Inspiration

Grief doesn't discriminate. One of the primary emotions that unite the world is grief. Medical grief and the loss of a loved one are hefty to move on with. One such crucial need is to know heart health, this web application helps the user predict the presence or absence of heart disease based on a few medical questions with answers depending on the user

What it does

This app helps the user sign up/login into our web application and provides resources about heart disease i.e how to prevent heart disease, causes of heart disease, and a test to check the presence of heart disease. This app shouldn't be mistaken for a professional medical diagnosis of any disease, please consult a doctor for the same.

How we built it

This web application was built using: ML algorithm - Random Forest Classifier. The application was mainly coded in python. Flask was used for the backend of the application. Sqlite3 for the database, email service - Gmail, deployed on - pythonanywhere.com, dataset - kaggle.com, and the frontend was made using - HTML, CSS, and Bootstrap.

Challenges we ran into

Identifying which data had the most impact on the prediction of heart disease and neglecting data that didn't seem to affect the result at all.

Accomplishments that we're proud of

This web Application is a solo major project and is successfully deployed, tested, and running on- http://10monica.pythonanywhere.com/

What we learned

That medical world is vast and there's always scope for more discoveries, more predictions, and anything that makes it feasible for people to access any sort of health service.

What's next for Heart Disease Prediction

Improved accuracy and UI. The Accuracy is mainly dependent on the random state we choose since the algorithm used was giving the highest accuracy, improved accuracy would make it easier for the user to believe and be satisfied with the prediction. Better UI for smoother function and attractive web application.

Share this project:

Updates