📰 Fake News Classification Project

A comprehensive machine learning project for detecting fake news using multiple algorithms and a user-friendly Streamlit web application.

🚀 Live Demo

Click here to view the deployed Fake News Classifier App

🎯 Project Overview

This project implements a complete fake news classification system with the following features:

• Multiple ML Models: Naive Bayes, Logistic Regression, KNN, SVM, Random Forest
• Text Vectorization: Both Count Vectorizer and TF-IDF Vectorizer
• Hyperparameter Tuning: Automated optimization for best performance
• Interactive Web App: Streamlit-based interface for real-time predictions
• Batch Processing: Support for CSV file uploads and bulk predictions
• Comprehensive Evaluation: Detailed performance metrics and visualizations
• Easy Deployment: Ready for local and cloud deployment

1 📁 Project Structure

Fake_News_Classifier/
│
├── data/ # Raw dataset
│ └── dataset.csv
│
├── models/ # Trained models & vectorizers
│ ├── best_model.pkl
│ ├── vectorizer.pkl
│
├── results/ # Model evaluation results
│ ├── confusion_matrix.png
│ └── model_comparison.csv
│
├── app/ # Streamlit application
│ └── app.py
│
├── preprocessing.py # Text preprocessing functions
├── model_training.py # Model training script
├── model_evaluation.py # Evaluation script
├── model_tuning.py # Hyperparameter tuning
├── prediction_pipeline.py # Prediction logic
├── requirements.txt # Python dependencies
└── README.md # You're here!

---

📊 Features

✅ ML Models Implemented

• Logistic Regression
• Multinomial Naive Bayes
• Support Vector Machine (SVM)
• Random Forest

🔤 Vectorization Methods

• CountVectorizer
• TF-IDF Vectorizer

🧪 Evaluation Metrics

• Accuracy, Precision, Recall, F1-score
• Confusion Matrix
• ROC Curve & AUC

🌐 Streamlit Web App

• Real-time prediction
• Batch prediction (CSV upload)
• Model performance view
• Probability & confidence visualization

---

⚙️ Setup & Usage

✅ Conclusion This Fake News Classifier project is a complete end-to-end system built with real-world practicality in mind. It combines robust machine learning models, clean text processing, and an interactive Streamlit interface — making it a great tool for both educational purposes and potential deployment.

By allowing both individual and bulk predictions with live confidence visualization, the system serves as a solid foundation for combating misinformation online.

🔮 Future Enhancements:

Deep learning models (BERT, LSTM)

Real-time news scraping & classification

Multi-language support

Browser extension integration

Thanks for visiting! 📰🚀