Data Mining and Machine Learning Techniques for Malware Detection

Abstract

Malware or malicious software is an umbrella term for viruses, worms, Trojans, spyware, and the like; it is a piece of code that is intentionally installed to infect computational devices. Several techniques have been proposed from time to time to detect these malwares. These techniques range from the early day signature-based detection to the machine and deep learning techniques. In the current scenario, the malwares use the techniques of obfuscation and polymorphism in order to hide themselves and go undetected. To detect these malwares, machine learning and data mining techniques are combined with existing detection methods in order to facilitate the detection process. Basic malware analysis techniques like static, dynamic, and hybrid have been detailed in this paper. In this paper, malware detection techniques have also been critically evaluated. This paper also focuses on the study of various data mining/machine learning approaches for malware detection proposed by different researchers.