An Algorithm for Monitoring Privileged Sessions and Detecting Risky Actions in Information Systems Based on Contextual and Behavioral Analysis
Year: 2026
Type: article
Abstract: Monitoring privileged user activity in information systems plays a vital role in protecting critical information resources. Conventional access control mechanisms verify granted permissions but cannot always detect their misuse. Effective privileged-session monitoring therefore requires joint assessment of habitual user behavior, sequences of performed actions, and associated risk levels. This paper proposes an integrated algorithmic model combining user profiling through K-Prototypes clustering, session anomaly detection using One-Class Support Vector Machines (OC-SVM), and contextual risk scoring. The mathematical formulation, operational steps, and experimental evaluation methodology of the proposed model are presented. The approach is designed to account for the relationships between contextual features and user actions within privileged sessions.
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7236983052
