📰 Fake News Detection Using NLP
📌 Project Overview
This project uses Natural Language Processing (NLP) to classify news articles as real or fake. The system achieves an impressive accuracy of 99.07% , making it highly reliable for detecting misinformation in real-world scenarios.
🎯 Key Achievements
Metric Score
Accuracy 99.07%
Precision 0.99
Recall 0.99
F1-Score 0.99
📊 Dataset
Feature Description
Source ISOT Fake News Dataset / Kaggle
Samples ~44,000+ news articles
Classes Real vs Fake News
Train-Test Split 80-20
🧠 Models Compared
Model Accuracy
Logistic Regression 95.40%
Random Forest 96.20%
XGBoost 97.10%
LSTM 98.50%
Bidirectional LSTM 99.07% 🏆
🛠️ Technologies Used
Python
Scikit-Learn
TensorFlow / Keras
NLTK
Pandas & NumPy
Matplotlib & Seaborn
📈 Key Features
✅ Text preprocessing (cleaning, stemming, stopwords removal)
✅ TF-IDF vectorization
✅ Multiple ML models comparison
✅ Bidirectional LSTM for deep learning
✅ Confusion matrix & classification report
📁 Project Structure
text
fake-news-detection/ │
├── fake_news_detection.ipynb # Jupyter Notebook
├── fake_news_detection.py # Python script
├── README.md # Documentation
├── requirements.txt # Dependencies
├── data/
│ └── fake_news_dataset.csv # Dataset
└── images/ ├── confusion_matrix.png └── gui_screenshot.png
🚀 How to Run
Clone the repository
Install dependencies
Download the dataset
Run the Jupyter Notebook
Launch the Gradio GUI
📊 Sample Output
text
✅ Accuracy: 99.07%
📊 Classification Report: precision recall f1-score support 0 0.99 0.99 0.99 4200 1 0.99 0.99 0.99 4400
🤝 Contributing
Contributions are welcome! Feel free to fork and submit a pull request.
📝 License
This project is licensed under the MIT License.
⭐ If you found this project useful, please consider giving it a star!
Let me know if you need any adjustments! 😊