MALWARE DETECTION USING MACHINE LEARNING TECHNIQUES

Abstract

Malware attacks have become serious and crucial issue now a days, as it can affect victim in many ways. Hence detecting malware at early stage is an essential aspect in the security of computer systems. Existing malware system contains a traditional antivirus detection method that depends on signature-based and behavioral methods. Traditional methods of malware detection are not that effective and cannot detect unknown malwares. In recent years machine learning is coming out as an emerging and challenging field in malware detection. Proposed method implements machine learning and deep learning technique for detecting malware. This is achieved using machine learning algorithm, Support Vector Machine and deep learning concept using Convolutional Neural Networks where in malwares are represented as images. The study compares the performance of conventional, machine learning-based, and deep learningbased malware detection techniques. Proposed method implemented for malware detection using Convolutional Neural Networks with malware images is more secure compare to dynamic based method as binary malware files are converted to images and images are never executed also it can reduce drawbacks of traditional signature based method at some extent.