Malware Detection Using Machine Learning

With the rise of polymorphic and zero-day malware, traditional signature-based detection methods have become insufficient. This paper proposes a hybrid malware detection framework that integrates machine learning (ML) and deep learning (DL) with dynamic behavioral analysis. By utilizing the CIC-MalMem-2022 dataset and extracting runtime features such as API calls and memory usage, the framework applies models like CNN, LSTM, and Decision Trees. Experimental results show that our approach achieves up to 99.27% accuracy while reducing false positives. This demonstrates the potential of intelligent, real-time, and scalable malware detection systems integrated with hardware-assisted techniques.