2026-08-15T13:38:29.817Z http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai

ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai:3:12142 2026-08-10T21:15:19Z 3

Anomaly Detection in Cloud Operation Logs Utilizing NLP and AdaBoost Tuned by Modified Crayfish Optimization Algorithm, Chapter in LNNS Lecture Notes in Networks and Systems: ICICC 2026: Innovative Computing and Communications, Springer, volume 2052 2026 http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai/record/3/12142 https://link.springer.com/chapter/10.1007/978-3-032-30909-9_30 S. Anetic M. Mihajlovic B. Radomirovic D. Bulaja T. Zivkovic M. Zivkovic L. Anicin N. Bacanin Robust log surveillance is critical for maintaining the efficiency and dependability of large-scale cloud computing infrastructures. Since these environments constantly expand in both intricacy and size, conventional monitoring strategies become increasingly inadequate and demand higher computational overhead. This study investigates the fusion of natural language processing methodologies with sophisticated AdaBoost classification model to detect anomalies within cloud-generated log records. Recognizing that a classifier’s precision is highly contingent on properly optimized hyperparameters, a metaheuristic-based optimization scheme is adopted to automate and enhance the tuning process. A novel adaptation of the crayfish optimization algorithm is devised and assessed using publicly accessible, real-world datasets. The experimental findings reveal impressive results, with top-performing models reaching accuracy rates near 98.5%, underscoring the method’s viability for intelligent, scalable anomaly detection in cloud-based logging systems. bookPart Springer, Cham 381 393 10.1007/978-3-032-30909-9_30 LNNS Lecture Notes in Networks and Systems: ICICC 2026: Proceedings of International Conference On Innovative Computing and Communications, volume 2052