🧠 Fake News Detector 📰

A smart, AI-powered app that verifies the authenticity of news using a custom-trained model combined with six real-world verification APIs. Built to tackle the challenge of misleading news by analyzing both linguistic patterns and contextual credibility.

---

🌟 Project Intuition & Motivation

Fake news isn't just misleading — it’s dangerous. Traditional ML models can classify text as fake or real, but news articles need context, language analysis, and external fact-checking to make an accurate judgment.

That’s why I went beyond just model training and built an integrated system that combines machine learning with 6 powerful APIs for real-time credibility analysis.

---

🚀 Approach & Architecture

1. Custom ML Model

   • Trained on over 50,000+ labeled news articles
   • Built using logistic regression, with preprocessing via NLTK and Scikit-learn
   • Achieved high accuracy (~98%) on test data

2. Layered Validation System

   • The app first checks the language and structural patterns using the ML model.
   • Then, it verifies the authenticity of the content by querying 6 fact-checking & credibility APIs:
     • NewsAPI
     • GNews API
     • Google Fact Check Tools API
     • MediaStack API
     • NewsData.io API
     • Currents API

3. Streamlit Web App

   • Clean, modern UI with real-time feedback
   • Dynamically styled feedback boxes (green/red)
   • Context-aware intermediate responses
   • Justified final verdict with AI-generated reasoning

---

🛠️ Tech Stack

• Python (NLTK, Scikit-learn, Pandas, Requests)
• Streamlit (for web app UI)
• APIs (for cross-verification)
• Google Colab (model training)
• Git + GitHub (version control)
• Mac/Linux & Windows support (via .sh and .bat scripts)

---

📊 Dataset Info

• Source: Kaggle “Fake and Real News Dataset”
• Data: ~50,000 articles
• Labels: Real vs. Fake
• Focus: Primarily political news (US-based)
• Preprocessing: Tokenization, stopword removal, TF-IDF

---

🔐 Why Use APIs?

“ML models learn language — not truth.”

Since fake news often includes real words with fake contexts, my app integrates external APIs to cross-check if the content is verified or disputed elsewhere.

This multi-layered pipeline ensures more realistic and trustworthy results.

---

📈 Future Improvements

• Auto-news summarization
• Language translation for Urdu & Hindi support
• Real-time news stream monitoring
• Deeper sentiment & emotion detection