Various Soft Computing Based Techniques for Developing Intrusion Detection Management System

Abstract

Protecting networks and data requires an effective Intrusion Detection System (IDS). Contextual knowledge processing may be used to identify attacks that are specific to certain applications and networks, which is becoming more common due to the fast advancement of network technology. A hybrid intrusion detection system may help overcome this problem (IDS). Packet flooding is a common tactic employed in DoS attacks, which aims to overload the victim's infrastructure. It is now possible to interrupt networks of virtually any size with these kinds of assaults. Testing a high-performance hybrid IDS is hampered by having to deal with massive amounts of data with many characteristics. In order to identify dangerous patterns, a high number of features might make it harder, leading to a protracted training and testing procedure, an increased resource demand, and a lower detection rate. As a result, minimizing and deleting irrelevant features from the benchmark dataset is needed. Preprocessing removes out-of-range values, unlikely data combinations, and missing values from the dataset. Feature Selection (FS) methods are used to reduce the dimensionality of a dataset by deleting irrelevant and redundant attributes, thus improving classifier accuracy. To solve this problem, the AODV protocol, in conjunction with soft computing techniques, is very useful.