2026-05-13T12:28:23.877Z http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai

ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai:3:10963 2025-01-26T13:57:24Z 3

Tuning Natural Language Processing by Altered Metaheuristics Algorithm for Phishing Email Identification, Chapter in LNNS Lecture Notes in Networks and Systems: CML 2024: International Conference on Computing and Machine Learning, Springer, volume 1144 2025 http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai/record/3/10963 https://link.springer.com/chapter/10.1007/978-981-97-7839-3_18 L. Jovanovic N. Bacanin R. Ravikumar M. Antonijevic M. Dobrojevic M. Zivkovic The high-speed development and integration of the Internet have led to the emergence of numerous technologies. Email has become an essential part of everyday life for most professionals. However, the human factor often remains the weakest link in the cybersecurity chain. Various actors seek to socially engineer their way into systems or mislead users into scams. Detection is often challenging, and traditional systems can be slow to adapt to changes in a dynamic system. Artificial intelligence is a promising approach for addressing phishing email detection in a dynamic environment using natural language processing. Detecting keywords and phrases in phishing emails may be an effective way to mitigate threats to systems and individuals. The proper application of AI algorithms requires careful hyperparameter selection. To tackle this issue, this work processes a modified version of a recently introduced algorithm to address hyperparameter selection. The approach introduced in this work is validated over a real-life dataset, and the results of the top-performing models exceed an accuracy of 90%, suggesting that effective detection of phishing emails can be handled using the introduced approach in the real world. bookPart Springer, Singapore 265 282 10.1007/978-981-97-7839-3_18 LNNS Lecture Notes in Networks and Systems: CML 2024: Proceedings of International Conference on Computing and Machine Learning, volume 1144