Fake News Detection using Machine Learning

Project Overview

Fake News Detection is a Machine Learning and Natural Language Processing (NLP) project designed to classify news articles as Real or Fake. The application analyzes textual news content, performs text preprocessing, extracts features using TF-IDF Vectorization, and predicts the authenticity of the news using a trained Machine Learning model.

The project aims to combat misinformation by providing an automated system capable of identifying potentially fake news articles in real-time.

Objectives

Detect whether a news article is Real or Fake. Apply Natural Language Processing (NLP) techniques for text analysis. Convert textual data into numerical features using TF-IDF. Train and evaluate Machine Learning classification models. Deploy the model using Streamlit for interactive predictions.

Technologies Used

Python Pandas NumPy Scikit-Learn Natural Language Processing (NLP) TF-IDF Vectorization Streamlit Pickle Git & GitHub

Project Structure

Fake_News_Detection/ │

├── app.py

├── Fake_News_Detection.pkl

├── requirements.txt

├── .gitignore

└── README.md

Workflow

1. Data Collection Collected news article datasets containing real and fake news records.
2. Data Preprocessing Removed unwanted characters and symbols. Converted text to lowercase. Removed stop words. Cleaned and prepared textual data for analysis.
3. Feature Engineering Applied TF-IDF (Term Frequency–Inverse Document Frequency) Vectorization. Converted text data into numerical vectors.
4. Model Training Trained Machine Learning classification models. Evaluated model performance using classification metrics.
5. Model Deployment Saved trained model using Pickle. Built a Streamlit web application for real-time predictions.

Features

1. Real-Time News Classification
2. User-Friendly Streamlit Interface
3. NLP-Based Text Processing
4. TF-IDF Feature Extraction
5. Machine Learning Prediction Engine
6. Fast and Accurate Predictions

Machine Learning Pipeline

News Article

  ↓

Text Cleaning

  ↓

Preprocessing

  ↓

TF-IDF Vectorization

  ↓

Machine Learning Model

  ↓

Real / Fake Prediction

Installation

Clone Repository

git clone https://github.com/Venky-eelaprolu/Fake_News_Detection.git

Move into Project Directory

cd Fake_News_Detection

Install Dependencies

pip install -r requirements.txt

Run Application

streamlit run app.py

Skills Demonstrated

Machine Learning

Natural Language Processing (NLP)

Text Classification

Data Preprocessing

Feature Engineering

TF-IDF Vectorization

Streamlit Deployment

Python Programming

Model Serialization (Pickle)

Project Outcome

Successfully developed a Machine Learning-based Fake News Detection system capable of classifying news articles as Real or Fake using NLP and TF-IDF Vectorization techniques. The application provides real-time predictions through an interactive Streamlit interface and demonstrates practical implementation of text classification models.

Future Enhancements

Deep Learning Models (LSTM, GRU)

Transformer Models (BERT)

News Source Credibility Analysis

Multi-Language Support

Explainable AI (XAI) Integration

Real-Time News API Integration

Author

Venkata Durga Rao Eelaprolu

🔗 GitHub: Venky-eelaprolu GitHub

🔗 LinkedIn: Venky Eelaprolu LinkedIn