🧠 Patient Disease Prediction System

A machine learning project that simulates and predicts disease progression in patients using a custom-built Random Forest algorithm — without relying on scikit-learn models.

The system:

• Generates realistic patient medical records (cholesterol, blood pressure, glucose, BMI, etc.)
• Predicts which disease a patient might develop after 1, 2, 5, and 10 years
• Allows you to update patient data and re-evaluate predictions
• Trains and reuses your own Random Forest model

---

🚀 Features

• 🩺 Generate large, realistic dummy datasets of patients
• 🤖 Custom Decision Tree & Random Forest implementation (no external ML libraries)
• 🕒 Predict diseases across multiple future years (1, 2, 5, 10)
• 🔁 Update patient information dynamically
• 💾 Save and reuse trained models to skip retraining
• 📊 Evaluate model performance and accuracy

---

📂 Project Structure

📦 patient-disease-predictor
├── backend/
│   ├── __pycache__/
│   ├── data/
│   ├── models/
│   ├── treeUtility/
│   │   ├── node.py
│   │   ├── decisionTree.py
│   │   ├── randomForest.py
│   ├── utility/
│   │   ├── generatePatientData.py
│   │   ├── patientManagementSystem.py
│   ├── venv/
│   ├── api_server.py
│   ├── main.py
│   ├── requirements.txt
│
├── data/
│   ├── patients.csv
│   ├── labels.csv
│
├── frontend/
│   ├── .next/
│   ├── app/
│   │   ├── dashboard/ 
│   │   │   ├── page.js
│   │   ├── patients/ 
│   │   │   ├── page.js
│   │   ├── home/ 
│   │   │   ├── page.js
│   │   ├── favicon.ico
│   │   ├── globals.css
│   │   ├── layout.js
│   │   ├── page.js
│   │
│   ├── node_modules/
│   ├── public/
│   ├── .gitignore
│   ├── eslint.config.mjs
│   ├── jsconfig.json
│   ├── next.config.mjs
│   ├── package-lock.json
│   ├── package.json
│   ├── postcss.config.mjs
│   ├── README.md
│
└── README.md

🧰 Tech Stack

• Python 3.9+

• NumPy

• Pandas

• Custom Random Forest & Decision Tree implementation

💡 Future Improvements

• Add time-series modeling for more accurate long-term predictions

• Include more features like family history, smoking status, and exercise level

• Integrate with a lightweight web UI

• Add model persistence (save and load trained forests automatically)

🧑‍💻 Author

Bao Hoang

• 🎓 Computer Science Student @ Christopher Newport University
• 📍 Yorktown, VA