Outside the Closed World: On Using Machine Learning for Network Intrusion Detection

Abstract

There are many threats on the internet, but botnets rank among the most significant. Network confidentiality, integrity, and availability is compromised as part of malevolent activities. Recently, the adoption of machine learning-based methods has been suggested as a trustworthy method of identifying and defending botnets. This study presents four distinct machine learning models: Decision Tree, Regression Model, Naive Bayes Model, and Neural Network Model. Studies have been conducted to detect and simulate botnets using the CTU13 and ISOT botnet traffic dataset. We employ a number of criteria, such as accuracy, precision, and sensitivity, to assess each model's performance in identifying both known and unidentified botnet traffic patterns. According to our research, our machine learning models can identify botnet traffic whether it originates from a known or unknown botnet, and there is a significant improvement in performance over other models.