Sistema para detecção de intrusão de botnets utilizando aplicações de machine learning

Abstract

Communication tools and the continuous advancement of the Internet have also resulted in the sophistication of tools and methods to carry out attacks against users and their computers, with features that facilitate criminal activities in the cyber environment. Among cyber threats, botnets have characteristics and advantages that have expanded their use in recent years, becoming a tool employed extensively by attackers to conduct attacks and gain control of various devices connected to computer networks. The way these threats behave and are updated brings several challenges to the intrusion detection area. In this paper, a study is presented on the application of machine learning techniques in detecting botnets by analyzing network traffic flows. The study aims to show how pattern classification techniques can be applied in intrusion detection systems to identify similarities between the infrastructure of botnets, where works in the literature were studied to address an application that aims to improve the problems related to the attribute selection steps and the data processing, crucial steps in machine learning models.