A Study on Self-Configuring Intrusion Detection Model based on Hybridized Deep Learning Models

In order to achieve the necessary security guarantee, Intrusion Detection Systems (IDSs) play a vital role in all networks and information systems worldwide. One of the best ways to prevent malicious assaults is to use an IDS. In order to keep up with attacker's constant evolution of attack strategies and discovery of new attack vectors, IDS must likewise advance by implementing more advanced detection techniques. New research in the deep learning sector, including intrusion detection, has been made possible by the enormous rise in data as well as the considerable advancements in computer hardware technology. Machine Learning (ML) techniques that are adapted on data learning representations include deep learning as a sub-field. This study analyses 25 research papers that discuss about machine learning techniques and intrusion detection models by highlighting numerous drawbacks and the advantages of the current methodologies, which provides the interpretation of the suitable method for most of the IDS systems. This case shows different ways to explore the journals, publication years, metrics, achievements of numerical evaluation techniques, and a number of variables support other factors. On the other hand, a technique analysis that considers the benefits and drawbacks of the methods is also presented. The relative assessment of the methodologies provides a thorough clarification of the proposed motivation.