Publication

Abstract : As the amount of information on the computer network has grown enormously, intrusion detection systems have faced more difficulty dealing with vast-dimensional information that includes unnecessary and duplicate information. It takes time and effort to detect threats, which raises the alarming rate accurately. To deal with the high-class imbalanced problem and high-dimensionality issue, this paper uses SMOTE-Pearson’s correlation coefficient (PCC) with variance threshold to reduce the high-dimension features and solve the imbalance class issue. It uses an edge-IIoT dataset that is pre-processed for the use of machine learning classifiers. The results are calculated using various performance metrics showing great accuracy, precision, recall, and Matthews correlation coefficient (MCC) enhancement.