Abstract

Nowadays, security on the internet is a vital issue and therefore, intrusion detection is one of the major research problems for networks that defend external attacks. Intrusion detection is a new approach for providing security in existing computers and data networks. An Intrusion Detection System is a software application that monitors the system for malicious activities and unauthorized access to the system. An easy accessibility condition causes computer networks vulnerable against the attack and several threats from attackers. Intrusion Detection System is used to analyze a network of interconnected systems for avoiding uncommon intrusion or chaos. The intrusion detection problem is becoming a challenging task due to the increase in computer networks since the increased connectivity of computer systems gives access to all and makes it easier for hackers to avoid their traces and identification. The goal of intrusion detection is to identify unauthorized use, misuse and abuse of computer systems. This project focuses on algorithms: (i) Concept Drift based ensemble Incremental Learning approach for anomaly intrusion detection, and (ii) Diversity and Transfer-based Ensemble Learning. These are highly ranked anomaly detection models. We study and compare both learning models. The Network Security Laboratory-Knowledge Discovery and Data Mining (NSL-KDD99) dataset have been used for training and to detect the misuse activities.