Akademik Veri Yönetim Sistemi
KARADENIZ 18TH INTERNATIONAL CONFERENCE ON APPLIED SCIENCES, Rize, Türkiye, 4 - 06 Temmuz 2026, ss.579-589, (Tam Metin Bildiri)
Cyberbullying has become an important social problem with the spread of digital technologies. Social media, messaging applications and online platforms pave the way for the easy spread of negative interactions between individuals. In this study, the definition of cyberbullying, its effects, detection methods and approaches to combating current technologies are discussed in a multidimensional manner. In the light of domestic and foreign studies in the literature, text mining and natural language processing (NLP) techniques developed for the analysis of cyberbullying content, especially in the Turkish language, have been examined in detail. In this context; keyword-based methods, sentiment analysis, machine learning algorithms, deep learning models (CNN, LSTM, BERT) and social network analysis have been comparatively evaluated. The study emphasizes that cyberbullying occurs not only through textual content but also through visual, video and user interactions and draws attention to the importance of detection systems to be developed with the integration of multiple data types. In addition, it has been emphasized that contextual analysis and preprocessing steps are decisive for success in additive and low-resource languages such as Turkish. Although deep learning models have been shown to achieve high accuracy rates, issues such as data diversity and ethical data sharing are still fundamental problems that need to be solved. As a result, early warning systems, algorithms that take cultural context into account, and real-time analysis capabilities should be developed to prevent cyberbullying.