Construction method of cyber attacks classifier on government information resources

• Main Authors: Володимир Леонідович Бурячок, Руслан Валентинович Грищук, Віктор Миколайович Мамарєв
• Format: Article
• Language: English
• Published: PC Technology Center 2015-01-01
• Series: Tehnologìčnij Audit ta Rezervi Virobnictva
• Subjects: information and telecommunication system; cyber attack; classifier; classification; decision trees; optimization
• Online Access: http://journals.uran.ua/tarp/article/view/37423
• ISSN: 2226-3780; 2312-8372

Dynamics of successful realizations of cyber attacks, the object of which are public information resources, demonstrates the need to improve their security. One problem that hinders the implementation of effective information security systems, such as attack detection systems, is their inability to provide reliable and timely event pattern classification of information and telecommunication systems. In research materials it is proposed an approach that enhances the efficiency of attack detection systems for government information resources by the speed criteria for the given parameters of classification accuracy. This is achieved through the introduction of CBA two-step classification scheme, based on binary grouping patterns of the system behavior. The developed construction method of cyber attacks classifier, based on decision trees and optimized flow of incoming data, can reduce the construction and operation of classification models at times and provides the performance of classification accuracy of system behavior patterns.