An Implementation of Novel Feature Subset Selection Algorithm for IDS in Mobile Networks

Intrusion identification and prevention in networks and network related sources has been considered as an all time challenge for researchers and network scientists. Research towards this is conducted using many data mining algorithms and ML techniques. Wide usage of Internet through hand held devices like mobiles and laptops which tracks users every step like location, activities, transfers etc., may be misused through various intrusions. The traditional security measures like firewall, authentication and cryptography used as first line of defence are not sufficient provided the full landscape of security challenges in the heterogeneous and distributed networks of the real world. Therefore, there is need for Intrusion Detection Systems (IDS) that are to be integrated with anti-virus and other security mechanisms to ensure end-to-end security. The existing IDSs suffer from the problem of massive amounts of network traffic that lead to unsatisfactory results or inaccurate intrusion results or computational difficulties that form barrier in providing timely protection to a communication network. It is important to have intelligent security systems that can segregate relevant features from irrelevant ones, discard irrelevant features, and remove redundant features besides discovering Indicative features to enhance the scalability and efficiency of IDS. This is the challenging problem that is addressed in this work which is aimed at proposing and implementing a comprehensive intrusion detection framework which not only takes care of detecting intrusions but also reduces feature space more elegantly and intelligently.