Phishing Detection System

Table of Contents

• Project Description
• Project Structure
• Dependencies
• Usage
• Model Selection and Evaluation
• About the Author
• License

Project Description

The Phishing Detection System is a web application that uses machine learning to predict whether a given URL is a phishing site. The application is built using Flask for the backend, with HTML5 and CSS for the frontend.

Project Structure

The project is structured as follows:

• static/: This directory contains static files like CSS and JavaScript.
• templates/: This directory contains the HTML templates.
• models/: This directory contains the trained model.
• app.py: This is the main script that runs the application.
• featureExtraction.py: This script contains the feature extraction logic.

Dependencies

This application requires the following Python libraries, which can be installed by navigating to the project directory and running pip install -r requirements.txt:

• Flask
• joblib
• numpy
• python-whois
• scikit-learn

Usage

You can run the application by executing the app.py script:

python app.py

This will start a local server and serve the Phishing Detection System on localhost:5000.

Model Selection and Evaluation

We initially tried using a Multilayer Perceptron (MLP) for this task, but it did not yield satisfactory results. We then switched to a RandomForest model, which significantly improved the performance of our phishing detection.

About the Author

Aditya Varma is a computer science graduate from the University of Wollongong. He has a keen interest in AI, cybersecurity, systems analysis, and web development.

License

This project is licensed under the MIT License - see the LICENSE file for details.