On the Assessment of Information Security Ensuring Models of Critical Information Infrastructure Facilities

Abstract

The main stages of the development of information protection systems at critical information infrastructure facilities were analyzed. The necessity of improving information protection systems against computer attacks was justified. The task of assessing the quality of computer attack detection models at the level of rules was presented. The problems of the classical approach were formulated while forming the rules for detecting computer attacks. An approach is proposed for solving the intrusion detection problem, based on machine learning. The requirements were formulated to eliminate the possible disadvantages of recognizing a computer attack. The advantages of the selected model were formulated. Graphs were constructed illustrating the asymptotic complexity of traffic classification approaches. The quality metric of the resulting model was developed. The model has been evaluated through cross- validation and its accuracy has been evaluated. The results were analyzed and the necessary conclusions were drawn about the sufficient effectiveness of the developed model.