Sensor Data Analysis and Management

3 Evaluation of Feature Selection Techniques in Intrusion Detection Systems Using Machine Learning Models in Wireless Ad Hoc Networks

T.J. Nagalakshmi¹, M. Balasaraswathi², V. Sivasankaran³, D. Ravikumar⁴, S. Joseph Gladwin⁵, S. Pravin Kumar⁶

¹ Assistant Professor, ECE, Saveetha School of Engineering, SIMATS, Chennai²  Associate Professor, ECE, Saveetha School of Engineering, SIMATS, Chennai ³  Associate Professor, ECE, SITAMS, Chithoor, AP⁴  Professor, ECE, Vel’s University, Chennai ⁵  Associate Professor, ECE, SSN College of Engineering, Chennai ⁶  AI Engineer, Smartail Pvt Ltd, Chennai.

3.1 Introduction

WANs operate with the same components as used in wired networks. But, in WANs, data are transferred over air medium. Hence, it has very poor security. In recent years, the growth of network-based service has been incredible. Therefore, network security becomes one of the important problems in cyber world. Bace and Mell [1] define intrusion as “efforts to compromise the confidentiality, dependability, and simplicity, or to avoid the security mechanisms of a computer or network.” Using this principle, an intrusion detection system (IDS) can be defined as a software or hardware used to detect efforts to compromise the privacy, reliability, or accessibility of a network, or to evade the security contrivances of a network [2].

Typically, multiple nodes exist in an area, and all nodes share the same wireless medium. Therefore, there are chances for data hacking or attacks. Several ...