Malware detection system using machine learning and DATA-mining techniques

Abstract

A serious threat these days is malicious executables. It’s designed to wreck computing system and a few of them cover network while not the information of the owner victimisation the system.iTwo approaches are derived for it i.e. Signature primarily based Detection and Heuristic primarily based Detection. These approaches performed well against celebrated malicious programs however cannot catch the new malicious programs. Totally different researchers have planned ways victimisation data processing and machine learning for police investigation new malicious programs. The strategy supported data processing and machine learning has shown sensible results compared to alternative approaches. This work presents static malware detection system victimisation data processing techniques like data Gain, Principal part analysis, and 3 classifiers: SVM, J48, and Naïve mathematician. For overcoming the dearth of usual anti-virus product, this paper has a tendency to use ways of static analysis to extract valuable options of Windows letter file as well as to extract raw options of Windows executables that area unit letter header data, DLLs, and API functions within every DLL of Windows letter file. Thereafter, data Gain, job frequencies of the raw options area unit calculated to pick out valuable set options, so principal part analysis is employed for spatial property reduction of the chosen options. By adopting the ideas of machine learning and data-mining, this research work constructs a static malware detection system that features a detection rate of 99.6%.