ISSN: 2265-6294

UNCOVERING NETWORK VULNERABILITIES THROUGH MACHINE LEARNING

Main Article Content

KAMARAPU NARESH ,VARUN MARAMRAJ

Abstract

Unlike earlier times, there have been notable transformations brought about by the progress made in personal computer and communication technology. While there are many advantages to embracing modern technology for individuals, organizations, and governments, some people are not so fond of it. For example, the availability of information, the security of information transfer methods, the protection of sensitive data, and so forth. Fear-based digital oppression is one of the main issues we are currently facing in light of these issues. Due to a number of groups, including the criminal underworld, professionals, and digital activists, digital dread—which has produced several issues for both persons and organizations—has grown to the point where it may jeopardize open and national security. Intrusion Detection Systems (IDS) were created as a result to avoid online attacks at any costs. Currently, port sweep efforts are distinguished by learning the Support Vector Machine (SVM) computations based on the new CICIDS 2017 dataset with 97.80%, 69.79% accuracy rates attained independently. Alternative algorithms that outperform SVM in terms of accuracy include random forest, convolutional neural network (CNN), and artificial neural network (ANN) (93.29, 63.52, 99.93, and 99.11, respectively)

Article Details