A Model for the Disclosure of Probe Attacks Based on the Utilization of Machine Learning Algorithms

The Probe attack is one of the most perilous kinds of cyberattacks since its primary objective is to get knowledge on the vulnerabilities of the target network. In the course of this research, an Intrusion Detection and Prevention Systems (IDPS) detection model that can identify Probe attacks will be developed using machine learning approaches. The proposed Probe attack detection model was examined by utilizing the NSL-KDD dataset in conjunction with the Support Vector Machine method, Naive Bayes method, Decision Tree method, Random forest method, Logistic Regression method, and K-nearest neighbor method. Matthews Correlation Coefficients (MCC), Recall, Precision and Accuracy are the metrics that are utilized in order to make a comparison between these six methods. In general, the performance of all methods is satisfactory when using the suggested model. On the other hand, the Random forest approach has shown the best performance across all six measures, while the Naive Bayes method gave the worst results possible.