ZeroDayHunter
A Machine Learning-Based Malware Detection System
ZeroDayHunter is a group senior project designed to detect malicious and benign files using machine learning models such as Random Forest and XGBoost.
View Project
A Machine Learning-Based Malware Detection System
ZeroDayHunter is a group senior project designed to detect malicious and benign files using machine learning models such as Random Forest and XGBoost.
ZeroDayHunter aims to improve malware detection by using machine learning techniques instead of relying only on traditional signature-based detection. The system analyzes file features and classifies samples as malicious or benign.
Traditional malware detection methods struggle to detect unknown, zero-day, and polymorphic malware. This project explores machine learning as a more adaptive approach for identifying threats based on extracted features and learned patterns.
This project was completed as a group senior project at the University of Bahrain, College of Information Technology.
The project was developed collaboratively, with each member contributing to specific areas to ensure successful completion.
The system loads data from the EMBER dataset, extracts features, handles missing values, and prepares the data for machine learning.
Machine learning models such as Random Forest and XGBoost are trained using processed training data.
The models are evaluated using accuracy, ROC-AUC, classification reports, and confusion matrices.
Random Forest Accuracy
XGBoost Accuracy
ROC-AUC Score
Email: zerodayhunter6@gmail.com
GitHub: Backend_ML_Malware-Detection-and-Analysis
GitHub: Frontend_ML_Malware-Detection-and-Analysis
GitHub: Malware-detection-ml