DDoS Attack Detection Using Machine Learning
1. Introduction
This project aims to develop an interactive web-based system for detecting Distributed
Denial-of-Service (DDoS) attacks using machine learning. The system is built using
Flask (Python) as the backend, HTML/CSS (Bootstrap) as the frontend, and a Naïve
Bayes classifier to predict potential DDoS attacks based on uploaded network traffic
data.
2. Technologies Used
• Programming Language: Python (Flask Framework)
• Frontend: HTML, CSS, Bootstrap
• Machine Learning: Naïve Bayes Classifier
• Dataset: APA-DDoS-Dataset 2.csv
• Libraries:
o pandas for data handling
o numpy for numerical operations
o sklearn for ML model training
o flask for web framework
o pickle for model persistence
3. System Architecture
1. Frontend: Allows users to upload network traffic data (CSV format).
2. Backend (Flask): Processes the file and sends it to the ML model.
3. Machine Learning Model: Predicts if the uploaded data contains a DDoS attack.
4. Result Page: Displays classification results.
4. Installation and Setup
4.1 Prerequisites
Ensure you have the following installed:
• Python 3.x
• Flask (pip install flask)
• Pandas (pip install pandas)
• Scikit-learn (pip install scikit-learn)
• Bootstrap (Included via CDN in HTML)
4.2 Running the Project
1. Clone the repository or place the project files in a directory.
2. Navigate to the project directory and run:
3. python ddos.py
4. Open your browser and go to http://127.0.0.1:5000/.
5. Upload a CSV file and click Detect Attack.
6. The system will display the result.
5. Implementation Details
5.1 Machine Learning Model
• Algorithm Used: Gaussian Naïve Bayes Classifier
• Training Data: Preprocessed APA-DDoS-Dataset
• Feature Engineering: Label encoding applied to categorical data
• Accuracy Achieved: ~95% on test dataset
5.2 Flask Backend
• Loads the trained model (ddos_model.pkl)
• Accepts user-uploaded CSV files
• Preprocesses the data and makes predictions
• Renders the appropriate result page (index1.html, index2.html, or index3.html)
5.3 Frontend
• HTML form for file upload
• Bootstrap-based UI for better responsiveness
• AJAX used for smoother interactions (optional enhancement)
6. Conclusion
This project successfully integrates a machine learning model into a web-based
system to detect DDoS attacks from uploaded network traffic data. Future
improvements can include deep learning models, real-time monitoring, and API
integration for live traffic analysis.
7. Future Enhancements
• Implementing real-time traffic analysis
• Adding visualizations for detected threats
• Deploying the system on a cloud server
8. References
• Flask Documentation
• Scikit-Learn Documentation
• Bootstrap Framework
