📰 Fake News Detector

A machine learning-powered web application that analyzes news articles to determine whether they are real or fake. Built with Python, Scikit-Learn, NLTK, and Streamlit.

Overview

In the era of rapid information sharing, fake news can spread quickly and cause significant harm. This project uses Natural Language Processing (NLP) and a Logistic Regression classifier trained on a dataset of over 20,000 news articles to predict the authenticity of a given news text.

Key Features

• Real-Time Analysis: Simply paste an article's text, and the model instantly predicts its authenticity.
• NLP Preprocessing: Robust text preprocessing including non-alphabetic character filtering, lowercasing, stemming (PorterStemmer), and stopword removal.
• TF-IDF Vectorization: Transforms textual data into meaningful numerical features for the ML model.
• Beautiful UI: A sleek, dark-themed responsive interface built with Streamlit and custom CSS.

Getting Started

Prerequisites

Make sure you have Python installed. You'll also need the following libraries:

• streamlit
• pandas
• scikit-learn
• nltk

Installation

1. Clone this repository to your local machine.
2. Install the required dependencies:

   pip install streamlit pandas scikit-learn nltk

3. Run the application:

   streamlit run app.py

4. Open the provided localhost URL in your browser.

How it Works

1. Input: The user pastes a news article into the Streamlit web interface.
2. Preprocessing: The text is cleaned. Special characters are removed, the text is converted to lowercase, and NLTK removes common English stopwords and applies Porter Stemming.
3. Vectorization: The cleaned text is transformed into a numerical format using the pre-fitted TfidfVectorizer (tfidf_vectorizer.pkl).
4. Prediction: The LogisticRegression model (fake_news_model.pkl) evaluates the vectorized text and returns a prediction (0 for Real, 1 for Fake).

Limitations & Future Work

Note: The current model was trained heavily on political news from the 2016 US election era. Due to the nature of machine learning, it may struggle to accurately classify short, out-of-context sentences, or news topics that differ vastly from its training dataset (e.g., international economic reports). Future versions will focus on retraining the model with a more diverse, generalized dataset.

License

This project is open-source and available for educational purposes.