Моделювання Причинно-Наслідкових Зв’язків У Безпеці Інформаційних Систем На Основі Нечітких Когнітивних Карт

Introduction. The main purpose of the developed model is adaptive approximation of cause-and-effect relationships between identified vulnerabilities, potential threat scenarios, and final consequences in the context of ensuring the functional security of modern information systems. The developed model aims to ensure effective processing of highly variable input data, robustness to noise, and dynamic forecasting of threat development in near real time. This is of crucial importance given the rapid growth in the complexity of information systems architecture. Materials and Methods. This study develops a neuro-fuzzy model based on fuzzy cognitive maps that combines Hebbian learning with error backpropagation in a hybrid training scheme. The proposed approach allows for simultaneous qualitative analysis of cause-and-effect relationships and high-precision numerical calibration of model parameters. Results. By accounting for hidden dependencies, the developed approach demonstrates significantly higher efficiency than conventional solutions, in particular Bayesian networks and static expert rules for identifying threats to functional security. Empirical data obtained during experimental research and modeling indicate the significant potential of the proposed method for flexible risk assessment and calibration. Conclusions. The model demonstrates high adaptability when applied in large-scale corporate networks, Internet of Things environments and critical infrastructure facilities. The presented concept is characterized by a strong self-learning capability, detection of atypical anomalous behavior and adaptation to new types of threats, resistance to information uncertainty and the possibility of integration into a wide range of existing information platforms.