Published Paper Details:

This survey study examines malware detection methods and focuses on a particular project that employs machine learning to find potentially hazardous URLs and identify infected files. Malware is any software that carries out harmful operations, such data theft or espionage. In Kaspersky Labs' definition from 2017, malware is "a type of computer programme designed to infect a legitimate user's computer and cause harm in various ways." The several polymorphic layers present in contemporary malware apps make it increasingly challenging to identify malware using conventional signature-based techniques. Since these layers either conceal from view or update themselves automatically on a regular basis, antivirus software finds it difficult to identify them. Machine learning provides the answer by teaching models to recognise both positive and negative file attributes. This feature allows malware to be detected independent of previous experience by enabling the recognition of hazardous patterns. The study that is being presented makes use of a Random Forest classifier for PE-header-based malware detection and Logistic Regression to discover potentially harmful URLs.