2026-05-12T21:58:37.605Z http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai

ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai:3:11470 2025-06-27T23:16:11Z 3

Social Media Harassment Detection Based on Modified Metaheuristic Optimized Adaboost Classification, Chapter in LNNS Lecture Notes in Networks and Systems: AITA 2024: Artificial Intelligence: Theory and Applications, Springer, volume 5588 2025 http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai/record/3/11470 https://link.springer.com/chapter/10.1007/978-981-96-1918-4_22 D. Bulaja L. Jovanovic J. Perisic M. Antonijevic M. Zivkovic N. Bacanin As society increasingly gravitates towards digital platforms for social interactions, the concerning rise in cyberbullying and harassment demands thorough examination. The migration of traditional social interactions to online spaces has unfortunately ushered in a surge of toxic behaviors, prominently among them being bullying. The cloak of anonymity provided by online platforms further complicates matters, rendering prevention efforts more challenging. This research delves into the potential of leveraging artificial intelligence (AI) classifier algorithms, empowered by sophisticated natural language processing (NLP) techniques, to not only detect but also mitigate instances of harassment. This inquiry is underscored by the urgent necessity to tackle this pervasive issue on a global scale. Additionally, the study introduces a refined metaheuristic optimizer, seamlessly integrated into the methodology, aimed at enhancing algorithmic performance. The methodology was evaluated on the Twitter harassment detection dataset. Comparative analysis demonstrated the superiority of models optimized through this novel approach, with an impressive accuracy rate of 0.829278, thereby accentuating its efficacy in combating online harassment. bookPart Springer, Singapore 315 328 10.1007/978-981-96-1918-4_22 LNNS Lecture Notes in Networks and Systems: AITA 2024: Artificial Intelligence: Theory and Applications, volume 5588