Using Order Texture Statistics to Classify Multi-Class Malware

The most significant challenge in the security of information and communication networks is the escalating number of malware types, followed by the search for appropriate methods to protect systems against them, which is one of the most critical concerns of programmers and information security specialists, along with the prompt recognition and identification of methods to combat malicious effects, such as malware. The purpose of this study is to use static and dynamic analytic approaches, as well as first- and second-order texture statistics, to detect and classify multi-class malware. The most important conclusions reached were that the order texture statistics approach provides better results than traditional methods (decision tree algorithm and naïve Bayes algorithm) in terms of improving precision, detection rate and false alarm rate, indicating the algorithm’s efficacy in cyberattack detection systems.

KEYWORDS

statistics approach, naïve Bayes, decision tree, cyberattack, static analysis, machine learning