Phishing URL Detection using Machine Learning

Overview

Phishing attacks remain one of the most common and effective cyber threats, often relying on deceptively crafted URLs to trick users into revealing sensitive information. This project implements an end-to-end phishing URL detection system using machine learning, exposed through a lightweight Flask web application.

Users can submit any URL and receive a real-time classification (legitimate or phishing) along with confidence scores.

---

Live Demo

Access the deployed application here:
https://phishing-url-detector-kn1b.onrender.com

Note: The application may take up to ~1 minute to load on first access due to cold start on free hosting.

---

Key Features

• Machine learning–based phishing detection using URL text only
• TF-IDF feature extraction at both word-level and character-level
• Linear Support Vector Classifier (LinearSVC) with calibrated probability outputs
• Confidence-aware predictions displayed through the web interface
• Flask-based web interface for real-time interaction
• Deployment-ready for platforms such as Render or GitHub-hosted environments

---

Dataset and Model

Dataset

• Labeled dataset of legitimate and phishing URLs
• Source: https://huggingface.co/datasets/pirocheto/phishing-url
  The dataset contains a diverse mix of legitimate and phishing URLs and is commonly used as a benchmark for URL-based phishing detection tasks.

Model Pipeline

• TF-IDF Vectorization
  • Word-level n-grams
  • Character-level n-grams (effective for obfuscated and shortened URLs)
• Classifier
  • Linear Support Vector Machine (LinearSVC)
• Calibration
  • Probability calibration applied to enable meaningful confidence estimates

The trained model is serialized and loaded at runtime for efficient inference.

---

Project Structure

phishing-url-detector/
├── app.py             # Main Flask server
├── templates/
│ └── index.html       # Frontend HTML
├── model/
│ └── model.pkl        # Trained machine learning model
├── helpers.py         # Model training and utilities
├── requirements.txt   # Python dependencies
└── README.md          # Project documentation

Getting Started

Prerequisites

• Python 3.8 or higher
• pip for dependency installation

Installation and Running

Clone the repository and install dependencies:

git clone https://github.com/yourusername/phishing-url-detector.git
cd phishing-url-detector
pip install -r requirements.txt

Run the web app:

python app.py

Then open http://127.0.0.1:5000 in your browser.