Network Malware Detection Using Deep Learning Network Analysis

Abstract

Malware, short for malicious software, is designed for harmful purposes and threatens network security because it can propagate without human interaction by exploiting user’s vulnerabilities and carelessness. Having your system regularly scanned for malicious software is essential for keeping hackers at bay and avoiding the disclosure of sensitive data. The major drawbacks are the rapid creation of new malware variants, and it may become difficult to detect existing threats. With the ever-increasing volume of Android malware, the sophistication with which it can hide, and the potentially enormous value of data assets stored on Android devices, detecting or classifying Android malware is a big data problem. Security researchers have developed various malware detection and prevention programs for servers, gateways, user workstations, and mobile devices. Some offer centralized monitoring for malware detection software deployed on many systems or computers. The purpose of this essay is to critically examine the research that has been done specifically on malware detection. This paper proposes the Anti-Virus Software Detection for Malware with Deep Learning Network (AVSD-MDLN) framework to explore the possible threats. The two methods help in finding the threats. Dynamic Analysis for the Detection of Spyware (DA-DS) framework is framed to detect malicious malware, while the other is for classifying Android malware which is helped out through the Category in an Ensemble (CE) method. Prior malware detection methods are compared with the results of the proposed method. According to the research findings, the proposed approach achieves a higher projected time (0.5 sec) and detection accuracy (97.47%) than the existing situation machine learning and deep learning methodologies. Performance, correlation coefficient, and recall rate all improved in the suggested framework. Likewise, the negative rate (MPR) and the positive rate (PPR) also improved.