Malware Detection Using Machine Learning Models
説明
The criminal mind of adversaries is the most dangerous threat to enterprises and organizations worldwide. The creation and dissemination of malware across the Internet have given attackers of all levels an edge when it comes to compromise modern computer systems. Among the most promising tools to defeat nefarious initiatives, machine learning algorithms, and more specifically classifiers, arise as key techniques to detect malware infection across computer systems and consequently immunize the system at early stages of the propagation. In this paper, we present a performance comparison between neural networks and logistic regression for the problem of malware detection. Results show that neural networks and logistic regression are able to detect a malware presence considering the memory usage and the CPU usage of healthy and infected computers with an accuracy of more than 90%.
