Fuzzy Clustering by Local Approximation of Memberships with Decision Tree based Trust Intrusion Detection System for Clustered Wireless Sensor Networks Kumar S. Nandha*, Dr. Malmurugan N.** *Professor/Vice Principal, Dhanalakshmi Srinivasan Engineering College, Perambalur, India **Professor/Director, Mahendra College of Engineering, Salem, India Online published on 14 October, 2016. Abstract Security of Wireless Sensor Network (WSN) is always a major thing as it has widespread application in most of the major domains such as battlefield surveillance, healthcare, etc. Basically there are three main components that deal with security of wireless sensor network, prevention, detection and mitigation. But it is very difficult to prevent wireless sensor network always from malicious attacks so it is always important to detect them as early as possible so that the proposed method can react to the attack not harm to wireless sensor network. In this work, a Fuzzy clustering by Local Approximation of Memberships (FLAME) with decision tree based Trust Intrusion Detection System (TIDS) proposed for clustered wireless sensor networks. Initially, it considered the trust which is distributed among some other factors such as energy, reliability and data. It derives and formulates trust such as direct trust and recommendation trust from these factors. Trust based recommendation is integrated with FLAME based verification method to classify normal and abnormal (attacker node and data) of the data sets and node in terms of malicious behaviour. Experimental results show that the proposed FLAME based TIDS achieved better performance compared than existing trust based intrusion detection systems. Top Keywords WSN, trust system, intrusion detection, energy, reliability, FLAME, decision tree, clustering, classification, malicious attack. Top |