🔍 Fake News Detection System

A modern web application leveraging deep learning to detect fake news from legitimate content with up to 90% accuracy.

preview.mp4

📋 Table of Contents

• Overview
• Features
• Tech Stack
• Project Structure
• Installation
• Usage
• API Documentation
• Model Architecture
• Performance
• Future Work
• Contributing
• License

🌟 Overview

This project implements a Fake News Detection System using a fine-tuned BERT architecture. The system analyzes news content and classifies it as either legitimate or fake news with high accuracy, helping users identify misinformation in digital content.

✨ Features

• High Accuracy Detection: Custom BERT model with 90% classification accuracy
• User-Friendly Interface: Modern React frontend
• API Integration: Well-documented FastAPI backend for easy integration
• Reproducible Research: Complete Jupyter notebook with model training process

🛠️ Tech Stack

Backend

• Model: Custom BERT architecture based on bert-base-uncased from Huggingface
• Framework: FastAPI
• Language: Python 3.8+
• ML Libraries: PyTorch, Transformers, Scikit-learn

Frontend

• Framework: React 18+ with Vite
• UI Components: Shadcn UI
• Styling: Tailwind CSS
• HTTP Client: Axios
• State Management: React Context API

Development & Deployment

• Version Control: Git

• Documentation: Jupyter Notebooks
• Package Management: npm, pip

📁 Project Structure

fake-news-detection/
├── README.md                             # Project documentation
├── LICENSE                               # Project License
├── backend/                              # FastAPI server
│   ├── app.py                            # API entry point
│   ├── model_download_helper.py          # Script to automatically download model files from cloud storage
│   ├── notebook/                         # Jupyter Notebook
│   ├── models/                           # ML model definitions
│   └── requirements.txt                  # Python dependencies
├── images/                               # Project Screenshots
│   └── screenshot.jpeg                   # UI Preview
└── frontend/                             # Vite-React application
    ├── src/
    │   ├── components/                   # UI components
    │   ├── pages/                        # Application pages
    │   ├── ui/                           # ShadCN UI
    │   ├── context/                      # React Context for Manual Theme Switching
    │   ├── lib/                          # TW Library
    │   ├── App.jsx                       # Application
    │   ├── App.css                       # Application CSS
    │   ├── index.css                     # Main CSS file
    │   └── main.jsx                      # Main React Component
    ├── package.json                      # JS dependencies
    └── tailwind.config.js                # Tailwind configuration

🚀 Installation

Prerequisites

• Python 3.8+
• Node.js 16+
• npm or yarn
• Git

Clone the Repository

git clone https://github.com/ItsTSH/Fake-News-Detection-System
cd fake-news-detection-system

Backend Setup

cd backend
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt

Frontend Setup

cd frontend
npm install  # or yarn install

💻 Usage

Start the Backend Server

cd backend
uvicorn app.main:app --reload

Start the Frontend Development Server

cd frontend
npm run dev  # or yarn dev

Access the application at http://localhost:5173

Main Endpoint

Endpoint	Method	Description
/predict	POST	Submit text for fake news analysis

🧠 Model Architecture

The fake news detection model is based on the BERT base architecture from Huggingface with the following modifications:

1. Pre-trained Base: bert-base-uncased from Huggingface
2. Layer Freezing: Most BERT layers are frozen during training for transfer learning efficiency
3. Custom Classification Head: Additional layers for the specific task of fake news detection
4. Hyperparameter Tuning: Optimized learning rate, batch size, and training epochs

For detailed architecture and training process, refer to the Jupyter notebook in the model/notebooks/ directory.

📊 Performance

• Accuracy: upto 90% on test dataset
• F1 Score: 0.86
• Precision: 0.91
• Recall: 0.91

🗃️ Dataset

• Dataset Used: Kaggle: fake-and-real-news-dataset

🔮 Future Work

• Support for multi-language fake news detection
• Support for news content along with the title
• Integration with browser extensions for real-time checking
• Explainable AI features to highlight suspicious text segments
• Model versioning and A/B testing infrastructure
• User feedback loop for continuous model improvement

👥 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add some amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

---

⭐ If you find this project helpful, please consider starring the repository! 🌟

Made with ❤️ by Tejinder Singh Hunjan