A Machine Learning Technique to Detect Malware

Abstract

Abstract: Organizations have been threatened by malware for a long time, but timely detection of the virus remains a challenge. Malware may quickly damage the system by doing pointless tasks that burden it and prevent it from operating efficiently. There are two ways to detect malware: the traditional method that relies on the malware's signature and the behavior-based approach. The malware's behavior is characterized by the action it conducts when active in the machine, such as executing the operating system functions and downloading infected files from the internet. Based on how it behaves, the suggested algorithm finds the virus. The suggested model in this study is a hybrid of Support Vector Machine and Principle Component Analysis. For real Malware, our suggested model obtained an accuracy of 92.70% during validation, with 96% precision, 96.32% recall, and an f1- score of .96