Enhanced intrusion detection system for wireless sensor networks using dynamic recursive feature selection and fuzzy temporal decision tree classification

Abstract

Wireless Sensor Networks are used in several applications such as Environmental monitoring, Healthcare Applications, Transportation and Logistics, Precision Medicine and Agricultures, Industrial Applications, Smart grids, Smart Cities and buildings, and many more. As the Wireless Sensor Networks Applications continue to grow, there are also security issues and challenges alongside that need to be addressed. Moreover, machine Learning adaptation is increasing in all the domains of business and technology including the security applications. Machine Learning provides the benefit of automation, better data handling and analysis, improvement in business process and saves human intervention and time in decision making. Wireless Sensor Networks are vulnerable for Security Attacks because of the nature of deployment of the application and hence it is necessary to handle the security attacks. Most of the Wireless Sensor Network Applications are installed in open terrain areas and they are susceptible for unauthorized and unauthenticated access. Also the limited and restricted storage, processing and power sources make the Wireless Sensor Network more complicated in implementing the security measures in these systems. newlineApart from Physical security of the sensor nodes, it is necessary to ensure the data and computing security of the system. Several conventional mechanisms such as encryption, security keys and firewall based security monitoring are not sufficient for advanced Wireless Sensor Networks now in use. In such a scenario, the use of Intrusion Detection System for Wireless Sensor Networks is an efficient mechanism for securing the Wireless Sensor Networks. However, the current Intrusion Detection Systems are not able to provide the required level of performance in the detection of the known and unknown security attacks because of their lack in dynamicity and automation. The current Intrusion Detection Systems are to be updated with the benefits of newline