Logical inference models based on fuzzy production rules for intelligent cyberthreat detection systems

Kostiuk, Yuliia та Skladannyi, Pavlo та Khorolska, Karyna та Bebeshko, Bohdan та Sokolov, Volodymyr (2026) Logical inference models based on fuzzy production rules for intelligent cyberthreat detection systems Fifth International Workshop on Conflict Management in Global Information Networks, 4281. ISSN 1613-0073

Текст Y_Kostiuk_P_Skladannyi_K_Khorolska_B_Bebeshko_V_Sokolov_4281_FITM.pdf Download (2MB)

Анотація

The article proposes an approach to building logical inference models in intelligent cyber threat detection systems based on fuzzy production rules. The main idea is to combine a classic rule-based mechanism (a system based on production rules) with Explainable Artificial Intelligence (XAI) methods, in particular SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME), to ensure transparency in decision-making. The developed model implements a fuzzy interpretation of knowledge about attack signs, allowing inference even in cases of incomplete or uncertain input data. The paper presents the structure of the production knowledge base, the defuzzification mechanism, and an example of applying the model to detect a phishing attack with a step-by-step logic for processing input information. A hybrid architecture is proposed that allows the combination of rule-based inference with Machine Learning (ML) methods, ensuring adaptability and increased accuracy. The modeling results demonstrate the effectiveness of this approach for cybersecurity tasks in dynamic environments. The proposed model can be integrated into existing information security systems, such as Security Information and Event Management (SIEM) solutions, Intrusion Detection Systems (IDSs), or cyber incident analysis platforms, to inform explainable decisions regarding the threat level. Modeling results on CICIDS-2017 and UNSW-NB15 showed an accuracy of over 95%, with support for interpretability of decisions through SHAP and LIME modules

Тип елементу :	Стаття
Ключові слова:	logical inference; fuzzy logic; fuzzy production rule; cyber threat; explainable artificial intelligence; XAI; SHAP; LIME; rule-based system; fuzzy inference system; FIS; SIEM; machine learning; decision making; intelligent system
Типологія:	Статті у базах даних > Scopus (без квартилю)
Підрозділи:	Факультет інформаційних технологій та математики > Кафедра інформаційної та кібернетичної безпеки ім. професора Володимира Бурячка
Користувач, що депонує:	Павло Миколайович Складанний
Дата внесення:	03 Лист 2026 22:05
Останні зміни:	03 Лист 2026 22:05
URI:	https://elibrary.kubg.edu.ua/id/eprint/60044

Actions (login required)

Перегляд елементу