Reconnaissance attack detection via boosting machine learning classifiers

Abstract

With the advancement of Internet technologies, network security concerns are growing exponentially. One of the most difficult issues of network security is keeping it safe. To detect and identify any malicious behavior the network, many security techniques were deployed. Intrusion Detection Systems (IDS) is one of the most frequent strategies for mitigating the effects of these attacks. Reconnaissance is a common attack in computer networks in which the attacker gathers as much information as possible about the target before conducting an attack. Machine Learning (ML) classifiers are commonly used to distinguish between normal and abnormal network traffic. In this paper, Raeconnaissance attacks detection is an exam with the following ML classifiers: Adaptive Boosting (AdaBoost), Gradient Boosting, cat Boosting, and eXtreme Gradient Boosting (XGBoost) to determine the most effective classifier in identifying Reconnaissance attacks. Evaluation metrics used are accuracy, precision, F-measure True Positive. The experiment on the UNSW-NB15 dataset shows that the cat Boosting classifier is superior to the XGBoost, AdaBoost and Gradiant Boosting. © 2023 Author(s).