Fake-News-Detection-Using-Machine-Learning

This project investigates how machine learning techniques can be used to detect fake news articles. It uses real and fake news datasets and walks through a full CRISP-DM-inspired data science process — including loading, cleaning, exploration, modeling, and evaluation, all within modular Jupyter/Colab notebooks.

Project Overview

This repository contains four notebooks, each representing a stage in the pipeline:

1. Data_Loader.ipynb
   Loads and combines real and fake news datasets, assigns unique identifiers, and stores the unified dataset in a SQLite database.

2. Data_Understanding.ipynb
   Conducts exploratory data analysis (EDA) using frequency analysis and word clouds to identify common tokens and patterns in fake vs. real articles.

3. Data_Preprocessing&Validation.ipynb
   Cleans the data, merges article title and body into a content column, and performs a stratified split into training and testing sets.

4. Modeling_Evaluation.ipynb
   Trains and evaluates multiple classifiers (Logistic Regression, Naive Bayes, Random Forest, and Linear SVM) using TF-IDF features and outputs comparative performance metrics.

---

Dataset Description

The dataset consists of two CSV files:

• Fake.csv: News articles classified as fake.
• True.csv: News articles classified as real.

Each file includes:

• title: Headline of the news article
• text: Body of the article
• subject: Category of the article
• date: Publication date

These are merged and labeled (1 for fake, 0 for real) during ingestion.

---

Technologies & Libraries Used

• Languages: Python 3
• Libraries:
  • pandas, numpy, os, sqlalchemy
  • sklearn for machine learning and evaluation
  • nltk for natural language processing
  • wordcloud, matplotlib, seaborn for visualization
  • sqlite3 for lightweight data storage

---

Machine Learning Models Evaluated

Each model is trained using TF-IDF vectorized text features:

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

Evaluation metrics include:

• Accuracy
• Precision, Recall, F1-Score
• Confusion Matrix
• Visual comparisons across models

---

Database Structure

The project uses a single SQLite database (fakenews.db) containing:

Table Name	Description
news_article	Combined dataset from both CSV files
train_articles	Preprocessed training set (80%)
test_articles	Preprocessed testing set (20%)

---

How to Run the Project

1. Clone this repository and upload the CSV files (Fake.csv, True.csv) to the appropriate path.
2. Open the notebooks in order:
   • Data_Loader.ipynb
   • Data_Understanding.ipynb
   • Data_Preprocessing&Validation.ipynb
   • Modeling_Evaluation.ipynb
3. Run all cells sequentially in each notebook.
4. Review metrics and plots to analyze model performance.

---

Sample Visuals

• Word Clouds for Fake vs. Real News
• Token Frequency Plots
• Confusion Matrices per Model
• Accuracy/F1 Comparison Chart

---

Authors

Alex McAnnally, Kyle Govender, Craig Runnels

This project was developed as part of the Advanced Business Analytics coursework at the University of Alabama at Birmingham.

---

Future Plans

As this project continues to evolve beyond the classroom, the following enhancements and features are planned for future development:

Dockerization

Package the entire project in a Docker container to ensure consistent environment setup, simplify deployment, and streamline collaboration.

Create a docker-compose configuration to separate model serving logic from the front-end UI, enabling modular architecture and easier maintenance.

Front-End Interface

Develop a user-friendly graphical interface using Flask, Streamlit, or a similar framework to allow users to input articles or headlines and receive real-time predictions.

Include visualization features to show confidence scores, model explanations (e.g., feature importance), and prediction history.

Model Improvements

Experiment with more advanced NLP techniques, such as transformer-based models (e.g., BERT, RoBERTa) to improve classification accuracy and robustness.

Fine-tune hyperparameters and expand the training dataset to better generalize across diverse sources and writing styles.

API Development

Expose a RESTful or GraphQL API to allow external applications to programmatically interact with the fake news detection model.

Implement rate-limiting, logging, and authentication for security and monitoring.

Logging and Feedback

Add backend logging of prediction requests (with user consent) for performance analysis and retraining.

Introduce optional user feedback mechanisms to flag incorrect predictions and continuously improve model performance.

---

License

MIT License

Copyright (c) 2025 Alex McAnnally

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.