A New Privacy Preservation Intrusion Detection (PPID) Techniques for Multiclass Attacks to Measure Its Reliability for Detecting Suspicious Activities

Abstract

There is currently no way that can secure Supervisory Control and Data Acquisition (SCADA) systems from invasions. This technology is not only capable of withstanding numerous types of attacks, but it also prevents the data from being exposed when it is processed by other applications, particularly Intrusion Detection Systems (IDS). Enterprises with mission-critical control environments can have their SCADA systems overseen. Ensuring the security of sensitive information becomes increasingly challenging when physical and digital systems are interconnected. As a result, privacy preservation approaches have been effective in securing private information and identifying harmful actions; yet, they fall short when it comes to detecting errors and determining the sensitivity percentage of data that is disclosed. In order to identify intrusion events and prioritise data, our recently developed Privacy Preservation Intrusion Detection (PPID) approach makes use of the correlation coefficient and Expectation Maximisation (EM) clustering methods. With the power system datasets for multiclass assaults, we test this technique's capacity to reliably detect suspicious activity. As shown above, the experimental findings demonstrate that the proposed strategy is more efficient and effective than three other methods that can be used with current SCADA systems.