Enhancing Network Security with Convolutional Neural Networks: An Anomaly-Based Intrusion Detection Approach

Abstract

The development of a robust network intrusion detection system (IDS) is important since a network infiltration by malevolent users could seriously disrupt networks. Models used to identify attacks on network infrastructures are the subject of intrusion detection. A key component of intrusion detection is anomaly detection, whereby deviations from typical behavior suggest the existence of attacks, flaws, problems, etc. that may have been purposefully or inadvertently caused. Numerous models have advanced in identifying emerging system risks since the introduction of anomaly-based intrusion detection systems. Currently, cybersecurity uses machine learning (ML) and deep learning (DL) models to detect anomalous intrusions. Creating an anomaly-based intrusion detection system (IDS) that can quickly identify and categorize different types of attacks is the goal of this project. Anomaly-based intrusion detection systems must be able to pick up on users’ or systems’ constantly shifting behavior. Packet behavior as parameters in anomaly intrusion detection is being experimented with in this paper. Enhancing the current Network Intrusion System is the aim of this research. The incapacity of certain internet security to automatically stop harmful attacks served as the catalyst for our effort. The suggested IDS learns the behavior of the system using a back propagation artificial neural network (ANN). Through the use of convolutional neural networks, this research improves the quality, convenience, and dependability of Network Intrusion Detection Systems in internet services. This creates a platform that allows users to share information and, consequently, reduces the amount of time they spend checking for numerous intrusion attacks.