AI-Based Heart Disease Risk Predictor

• 

  AI correctly identified 161 out of 184 patients. Overall accuracy: 87.5%

• 

  SHAP Explainable AI showing which factors cause heart disease most. ST_Slope is #1 risk factor.

• 

  Dataset shows 410 healthy and 508 sick patients. Well balanced for AI training.

• 

  Random Forest feature importance showing ST_Slope, ExerciseAngina and Oldpeak as top factors.

• 

  Older patients (avg 57) have more heart disease risk. Cholesterol shows wider variation in sick patients.

Inspiration

Cardiovascular disease is the #1 cause of death globally. I wanted to build an AI tool that helps doctors detect heart disease risk early and save lives.

What I Learned

• How to build a Machine Learning model
• How to use Explainable AI (SHAP)
• How to analyze real medical data
• How Random Forest Classifier works

How I Built It

1. Used a dataset of 918 patients with 16 medical features
2. Cleaned and analyzed the data
3. Trained a Random Forest Classifier (87.5% accuracy)
4. Added SHAP for Explainable AI
5. Built a Patient Risk Predictor

Challenges I Faced

• Understanding medical data columns
• Fixing errors in the code
• Making AI explainable for doctors

Results

• Model Accuracy: 87.5%
• Precision: 0.87
• Recall: 0.87
• F1-Score: 0.87
• Most Important Factor: ST_Slope (ECG pattern)

Built With

• google
• matplotlib
• pandas
• python
• scikit-learn
• seaborn
• shap