FORMAL MODEL OF COMPUTER ATTACKS, BASED ON SYSTEM EVENT RECORDS OF THE OPERATING SYSTEM

Modern computer attacks are characterized by the complexity of detection, the combination of various techniques and the active use of new technologies by attackers, which requires the development of intelligent approaches to their detection. The purpose of this work is to create a formal model that makes it possible to identify signs of computer attacks based on the analysis of operating system system events. As part of the practical application of the developed method, a marked-up dataset was obtained. Various machine learning algorithms have been used to detect computer attacks. A comparative analysis showed that the «Random Forest» algorithm demonstrates the best result. The accuracy of the developed classifier was 99,65 %. The results obtained confirm the high efficiency of using the formal model in practical activities for detecting computer attacks