❤️ Heart Disease Prediction using Random Forest

📌 Project Overview

This project predicts the likelihood of heart disease using patient medical attributes.

The goal is to build a simple, interpretable, and generalized machine learning model that can assist in early risk detection.

🚀 Tech Stack

• React / Vite
• Flask API
• Python
• Scikit-learn
• Pandas
• Random Forest

📊 Dataset Features

The model was trained on 12 medical features:

• Age
• Sex
• Chest Pain Type (cp)
• Resting Blood Pressure (trestbps)
• Cholesterol (chol)
• Resting ECG (restecg)
• Maximum Heart Rate (thalach)
• Exercise Induced Angina (exang)
• ST Depression (oldpeak)
• Slope of the Peak Exercise ST Segment (slope)
• Number of Major Vessels (ca)
• Thalassemia (thal)

Target:

• 0 = No Heart Disease
• 1 = Heart Disease

🧠 Model Used

• Random Forest
• Class-weight balanced to handle dataset distribution
• Generalization through scaling and regularization

📈 Model Performance

• Training Accuracy: ~96%
• Test Accuracy: ~91%
• Evaluated using:
  • Accuracy
  • Confusion Matrix
  • Classification Report

The model was kept simple to avoid overfitting on a small dataset.

🎯 Why Random Forest?

• Excellent performance on structured tabular data
• Robust to outliers
• Captures non-linear relationships better than simple linear models
• Provides probability scores for risk assessment

🖥️ React & Flask App

An interactive web app allows users to input patient details and receive:

• Predicted risk
• Probability score

To run locally:

1. Start the Flask Backend:

python api.py

2. Start the React Frontend:

cd frontend
npm install
npm run dev