Security Improvement Malware Detection Using Impact of PCA Function Extraction

  • T. S. RajaRajeswari , P. Subhashini , S. Spandana

Abstract

Ransomware is one of today's most severe Internet security challenges. Indeed, most Internet issues, including spam e-mails and denial of service attacks, are triggered by malware. In many words, machines that are infected by ransomware are also networked into botnets, and often assaults are performed on hostile, assaulting networks. The installation of files such as.exe,.bat,.msi etc. from untrusted internet sites may be likely to contribute to maladministration. These executables are changed intelligently to circumvent antivirus specifications by anomalous users. In this article, an improved identification approach for harmful executables is suggested by evaluating Portable Executable (PE) executable files and utilizing an extraction process for PCA functions. We also learned a supervised binary classifier using these features from regular and malicious PE data. We have checked our system on a comprehensive publicly accessible dataset and obtained a rating accuracy of over 96%.

 

Published
2020-12-30
Section
Articles