🔍 Fake News Classifier with Explainable AI

The Fake News Classifier is a machine learning-based system designed to automatically detect misleading or false news articles. Built using natural language processing (NLP) techniques and classification algorithms, the project focuses on distinguishing between real and fake news headlines or articles, specifically within the Indian context.

✅ Key Features:

• Text Preprocessing: Tokenization, stopword removal, TF-IDF vectorization for robust feature extraction.

• Model Training: Trained multiple classification models like Logistic Regression, Random Forest, and Naive Bayes, achieving high accuracy on labeled datasets.

• Explainability: Integrated SHAP (SHapley Additive exPlanations) to provide interpretable model predictions — users can understand which words influenced a fake or real prediction.

• Dashboard & Insights:

  • Visual display of fake news trends over time
  • Top indicative words for fake and real news
  • Model confidence visualization

🧠 Tech Stack:

• Python, scikit-learn, Pandas, Matplotlib, SHAP, Streamlit (optional for dashboard)
• Dataset: Cleaned and balanced dataset containing labeled Indian news articles

🔍 Use Cases:

• Helping readers and researchers detect misinformation
• Educational tool to demonstrate how AI can be used responsibly in combating fake news
• Basis for further extensions into multilingual or social media fake news detection