Heart disease prediction model

Correlation Matrix of various variables
Accuracy score of various models that ran over various algorithms
Classifcation report of Decision Tree Classifier that has been chosen as the main model for this project

Inspiration

A lot of people suffer from heart related issues or complications, especially the old people. To prolong their life, we can help them by identifying if they are prone to heart related issues or not.

What it does

It takes various values such as heart rate at rest, blood pressure of the person along with other parameters totalling to 14, and predicts how likely someone is to face heart related issues.

How we built it

I used various ensemble based Machine Learning models such as Decision Tree, Random Forest, Adaptive Bossting and also have utilized Gradient Boosting and Logistic Regression . The final output is based on the AdaBoost model.

Challenges we ran into

Initially, the accuracy score of the model was around sub-60s and we used hyperparameter tuning to take it to 0.75.

Accomplishments that we're proud of

Getting a True Positive Rate(TPR) of 0.97 for positive cases which means that our model is able to find with high accuracy people who have heart ailments. This was achieved in Decision Tree Classifier.