Enhanced Elman spike neural network based intrusion attack detection in software defined <scp>Internet of Things</scp> network
P. Ravi Kiran Varma, R. R. Sathiya, M. Vanitha
Abstract
Summary In this article, enhanced Elman spike neural network based intrusion attack detection in software defined IoT network is proposed. Initially, the data's are taken from CICDDoS 2019 and CICIDS 2018 benchmark datasets. Software defined network (SDN) secure defense system is detected the intrusion and distributed denial of service (DDoS) attacks on central controllers using multidimensional internet protocol (IP) flow analysis. Here, the enhanced Elman spike neural network (EESNN) classifies DDoS and intrusion attacks as normal and anomaly. The proposed EESNN‐IAD‐SDN method is executed in python language. The performance metrics, such as accuracy, specificity, F‐measure, sensitivity, precision, recall is examined. The proposed EESNN‐IAD‐SDN method provides 13.93%, 13.26%, 14.35, and 13.73% higher accuracy in CICDDoS 2019 dataset compared with the existing methods, like GRU‐IAD‐SDN, LSTM‐IAD‐SDN, and GAN‐IAD‐SDN, respectively.