Malware Detection Using Machine Learning Techniques

Abstract

AbstractMalware detection and identification are the primary factor for many enterprise applications to provide security to data of an organization and end-to-end monitoring of the resources accessed by various users globally. Every application is accessed through the Internet and pervasive computing environments. Unauthorized users like intruders and malicious users are primarily responsible for accessing the user’s data through various methods to infiltrate or produce damage to the computer system from remote locations. Without any owner consent, the malware is a kind of computer-based threat which is the reason for file infectors and standalone malware across the distributed computing environments. This topic mainly focuses on detecting various types of malware based on a dataset and understanding the features and key attributes of the dataset in providing training based on the most advanced machine learning algorithms to identify various types of malware like adware, spyware, backdoors, worms, and another type of malware resource which infect the system in many numbers of ways. The idea behind the proposed methodology is to work on the dataset and identify the malware characteristics and features to extract based on the dataset and identify the various class labels to train a model and identify the malware based on the provided files. Finally, the accuracy of various machine learning algorithms is checked and produced results based on the classification and prediction of malware across the small- and large-scale datasets.KeywordsMalwareMachine learningData scienceTraining the modelsTesting the modelsComparative studyPrediction and detection-based methodsConfusion matrixAccuracy results