An optimal cluster & trusted path for routing formation and classification of intrusion using the machine learning classification approach in WSN

Generally, wireless sensor networks (WSN) are being utilized in a wide range of fields like queue tracking, military applications, environmental applications, and so on. This approach is an attempt to focus on the detection of attack with the utilization of machine learning and optimization strategies. Primarily, the system model is initiated and the nodes are deployed randomly based on the size of the network. The cluster formation will be carried out with the use of energy competent Particle swarm optimization depending on the passive clustering mechanism (ECPSO-PCM) strategy. Using spatial correlation, groups correlation group will be formed. The probability of transmission is then estimated by taking into account the spatial correlation, quality of link among CH and cluster member nodes, and the node's residual energy of the network. The management of the trust is employed by the selection of cluster heads. If node consists of the criteria for trust coverage, then this node is chosen as the cluster head. If this condition is not satisfied, then it is chosen as a cluster member. The optimal range of cluster paths for effective transmission of data is carried using the Computation of optimal cluster path using Bio-inspired Hierarchical order chicken swarm optimization (BIHO-CSO) at which the distance and residual energy are major constraints. Once the optimum and trusted path is chosen, the classification and detection of attack are carried out using a Recursive Binary partitioning decision tree classifier (RBP-DT). The performance analysis is made and the attained outcomes are compared with traditional approaches to validate the supremacy of the presented scheme