Selection of Machine Learning Methods for Keylogger Detection Based on Network Activity
Dmitry Sergeevich Levshun, Diana Levshun · 2024
A keylogger is a technology that secretly tracks user input into a system. Attackers use it to steal personal information and credentials, such as logins and passwords on banking websites and electronic payment systems. Therefore, detecting keyloggers is one of the priority tasks for ensuring information security. This paper presents an approach to detecting keyloggers based on a set of artificial intelligence methods that make it possible to identify implicit patterns. We are also experimentally evaluating machine learning models using the same software, hardware and the same subsets of the Keylogger dataset for training and testing. We compare ten intelligent models, including gradient boosting, decision trees, and k-nearest neighbors. This allows us to determine the most suitable keylogger detection models in a single environment.