2026-07-19T15:38:11.154Z http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai

ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai:3:12103 2026-07-12T00:58:26Z 3

Cardiovascular Conditions Risk Assessment Utilizing SMOTE Preprocessing XGBoost Optimized by Modified Firefly Algorithm, Chapter in LNNS Lecture Notes in Networks and Systems: ICDPN 2025: Data Processing and Networking, Springer, volume 1934 2026 http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai/record/3/12103 https://link.springer.com/chapter/10.1007/978-3-032-23297-7_40 N. Bacanin J. Petrovic P. Dabic T. Zivkovic M. Zivkovic B. Radomirovic I. Kosta L. Anicin Precise estimation of the patient’s susceptibility to various heart-related conditions is crucial to properly classify cases and ensure that medical care is timely and appropriate. This study responds to this demand by forecasting these risks in cardiology departments, utilizing a finely tuned machine learning framework. The categorization process used the XGBoost method, with its parameters systematically adjusted through a tailored adaptation of the firefly optimization algorithm. The methodology was assessed on a real-world dataset, preprocessed using the SMOTE technique to mitigate significant class imbalance. The developed system demonstrated strong dependability in anticipating cardiovascular risks, with the best performing models attaining close to 95.04% precision. These results emphasize the potential of the suggested strategy as a practical decision-supporting instrument, capable of accelerating clinical interventions and enhancing hospital resource management and distribution in the future. bookPart Springer, Cham 478 494 10.1007/978-3-032-23297-7_40 LNNS Lecture Notes in Networks and Systems: ICDPN 2025: Proceedings of International Conference on Data Processing and Networking, volume 1934