A Survey on Machine Learning-based Detection and Classification Technology of Malware

Under the background of massive malware, traditional methods of malware analysis can no longer efficiently meet the task of malware analysis. The development of machine learning technologies has effectively solved this problem. In recent years, machine learning and deep learning technologies have been gradually applied to the detection of malicious software. Machine learning-based malware classification has become a research hotspot in the field of malware detection. As an overview of machine learning-based detection and classification of malware, we analyze and summarize the system framework of malware detection and classification, introduce the basic categories of malware features and main classifiers of malware detection and classification, and propose a basic framework and technical architecture of feature processing technologies. Then, we summarize some issues and challenges faced by the current malware detection-related work. Finally, we discuss some potential research directions of machine learning-based detection and classification of malware from three aspects.