Method of Identifying the State of Computer System under the Condition of Fuzzy Source Data
Вантажиться...
Дата
2020
ORCID
DOI
Науковий ступінь
Рівень дисертації
Шифр та назва спеціальності
Рада захисту
Установа захисту
Науковий керівник
Члени комітету
Видавець
Анотація
The purpose of this work is developing a method for identifying the abnormal state of a computer system based on the Bayes' Fuzzy classifier. It allowed us to create a Fuzzy expert identification system with an unlimited number of controlled indicators that belong to a finite interval. Estimation of informativeness of such indicators does not depend on the type of indicator’s functions and on the rule of their usage in the calculated formula. Introduced criterion allowed to estimate indices of the functioning of computer systems presented indistinctly. The quality of classification was evaluated based on ROC analysis. It was found that the method based on Bayes' Fuzzy expert system is qualitative, and its classification speed is almost independent of quantity indicators. Comparative evaluation of Bayes' Fuzzy classifier with Fuzzy clustering classifier and Fuzzy discriminant classifier are performed. In order to regulate the level of false-positive and false-negative classification, recommendations have been developed to manage the level of sensitivity and specificity of a Fuzzy expert system based on the Bayes classifier.
Опис
Ключові слова
state identification, mission critical computer systems, identification measurements, identification of misuse
Бібліографічний опис
Gavrylenko S. Method of Identifying the State of Computer System under the Condition of Fuzzy Source Data [Electronic resource] / S. Gavrylenko, V. Chelak, M. Kazarinov // International Journal of Computer Science and Security (IJCSS). – Electronic text data. – 2020. – Vol. 14, iss. 5. – P. 174-186. – URL: https://www.cscjournals.org/manuscript/Journals/IJCSS/Volume14/Issue5/IJCSS-1583.pdf, free (accessed 19.09.2023).