Nephosense AI

Predicting and Managing Chronic Kidney Disease (CKD) Using AI

Nephosense AI is an advanced AI-powered healthcare application designed to predict the risk of Chronic Kidney Disease (CKD) and provide personalized treatment recommendations based on patient data. It leverages machine learning, clustering, and reinforcement learning to improve CKD diagnosis and management.

---

🚀 Features

🔍 CKD Prediction & Future Risk Analysis

• Patients provide test results (e.g., GFR, Serum Creatinine, BUN).
• Lifestyle factors like diet, physical activity, and smoking habits are collected.
• The system checks for CKD diagnosis based on test results.
• If no CKD is detected, it predicts:
  • Future values of GFR, Serum Creatinine, etc.
  • Risk of CKD diagnosis in the next year.
  • Approximate time when CKD might develop.

🩺 Personalized Treatment Plans

• Uses clustering (KMeans) to group similar patients.
• A supervised learning algorithm ranks treatment plans based on efficiency.
• Top 3 recommended treatment plans are displayed to the user.
• Patients can provide feedback, and a reinforcement learning model (Q Learning) adjusts the treatment plan and saves a modified version.

---

🛠 Tech Stack

• Frontend: React (Tailwind CSS for UI)
• Backend: Flask (Python)
• Database: PostgreSQL
• Machine Learning: Scikit-learn, XGBoost, Q-Learning

---

🔧 Setup & Installation

Backend (Flask) Setup

# Clone the repository
git clone https://github.com/yourusername/Nephosense-AI.git
cd Nephosense-AI/backend

# Create a virtual environment and install dependencies
python -m venv venv
source venv/bin/activate  # On Windows use: venv\Scripts\activate
pip install -r requirements.txt

# Start the Flask server
python app.py

Frontend (React) Setup

# Navigate to the frontend directory
cd ../frontend

# Install dependencies
npm install

# Start the React app
npm start

---

📊 Database Schema (PostgreSQL)

• Patients Table: Stores patient details, test results, and lifestyle data.
• CKD Predictions Table: Stores past CKD predictions and risk assessments.
• Treatment Plans Table: Stores predefined and dynamically generated treatment plans.
• User Feedback Table: Stores feedback for reinforcement learning adjustments.

---

📜 Machine Learning Approach

1. Clustering (DBSCAN): Groups patients based on similar medical profiles.
2. Supervised Learning Model: Ranks treatment plans for each patient group.
3. Reinforcement Learning: Improves treatment recommendations based on patient feedback.

---

🛡 Future Enhancements

• Adding real-time monitoring for kidney health.
• Integrating IoT wearable data for better tracking.
• Expanding the dataset for more accurate predictions.

---

🤝 Contributing

1. Fork the repository.
2. Create a new branch (git checkout -b feature-name).
3. Commit changes (git commit -m "Added new feature").
4. Push to the branch (git push origin feature-name).
5. Open a Pull Request.

---

.