This project implements a robust machine learning model designed to accurately detect phishing URLs by analyzing a diverse set of URL and website features. The goal is to enhance cybersecurity by identifying potentially malicious URLs before they can cause harm.
The input dataset (located in the valid_data/ directory) contains engineered features representing various characteristics of URLs. Each feature is encoded as:
1→ Positive or benign attribute0→ Neutral or unknown attribute-1→ Negative or suspicious attribute
These feature vectors are used by the model to classify each URL as:
1→ Phishing0→ Legitimate
- Load and preprocess the validated URL feature dataset i.e. valid_data/test.csv
- Use a trained machine learning model to assess phishing risk.
- Saves predictions to
predicted_output/output.csvfor review and further analysis.
- ✅ Accuracy: ~92%
- 📌 Precision and Recall: Over 90%
- 🔍 Key Predictors: SSL certificate status, URL length, domain registration length
- 📦 Handles a broad spectrum of phishing tactics for generalized robustness
This model serves as an automated solution for phishing URL detection, enabling proactive defense in cybersecurity systems.
The model and application were containerized and deployed using:
- Docker: For consistent containerized environment
- AWS ECR: To store and manage Docker images
- AWS EC2: As the hosting server to run the container in production
-
📥 Clone the repository :
git clone https://github.com/iam-salma/CrisisAid-news-and-awareness-website.git cd CrisisAid-news-and-awareness-website -
🐍 Make sure you have Python 3 installed. :
Here’s the official link to install Python 3: 🔗 https://www.python.org/downloads/
-
📦 Create a virtual environment :
python -m venv venv
-
⚙️ Activate the virtual environment
On Windows :
.\venv\Scripts\activate
On macOS/Linux :
source venv/bin/activate -
📌 Install dependencies :
pip install -r requirements.txt
-
🗝️ Create .env folder to store secrets :
store your MONGO_DB_URL
-
🏃To Run the Project:
python main.py
ENJOY 😊🎉