A novel two-phase cycle algorithm for effective cyber intrusion detection in edge computing

Abstract Edge computing extends traditional cloud services to the edge of network, closer users, and is suitable for network with low latency requirements. With rise computing, its security issues have also received increasing attention. In this paper, a novel two-phase cycle algorithm proposed effective cyber intrusion detection in based on multi-objective genetic (MOGA) modified back-propagation neural (MBPNN), namely TPC-MOGA-MBPNN. first phase, MOGA employed build optimization model that tries find Pareto optimal parameter set MBPNN. The applied simultaneous minimization average false positive rate (Avg FPR), mean squared error (MSE) negative true TPR) dataset. second some MBPNNs are created obtained by trained search more locally. phase used as input training process repeated until termination criteria reached. A benchmark dataset, KDD cup 1999, demonstrate validate performance approach detection. can discover pool MBPNN-based solutions. Combining these MBPNN solutions significantly improve performance, GA combination. results show achieves an accuracy 98.81% 98.23% outperform most systems previous works found literature. addition, generalized classification applicable problem any field having multiple conflicting objectives.