Mining of intrusion attack in SCADA network using clustering and genetically seeded flora-based optimal classification algorithm

The applications such as the remote communication and the control system are in critically integrated arrangement. The controlling of these network is specified by supervisory control and data acquisition (SCADA) systems. This study discusses about the attack prediction and classification process by using an enhanced model of machine learning technology. The attack types are classified by the optimal selection of features extracted from the sensor data. In this, the features are labelled and cluster between the matrixes are extracted. These cluster forms the initial processing of attack identification which prevents the mismatched result. This clustering of data is performed by mean-shift clustering algorithm. From that clustered data, the features that are irrelevant for classification process is identified and suppressed by using the genetically seeded flora optimisation algorithm. In this optimisation process, the flora seeds are selected genetically to select best features. Then, from that optimally selected clustered data, the relevancy vector is predicted and the types are classified. The classification process is performed by the Boltzmann machine learning algorithm. The classified results of the proposed method for testing SCADA dataset are analysed and the performance metrics are evaluated and compared with the state-of-the-art methods.