Metrics
1 Record Views
Conference proceeding
Analysis and Experimental Comparison of State-Of-The-Art Deep-Learning Classification Techniques for Cyberbullying Detection
Procedia computer science, Vol.246, pp.3800-3809
International Conference on Knowledge Based and Intelligent information and Engineering Systems (KES 2024), 28th (2024)
2024
Abstract
Cyberbullying is becoming a relevant challenge in next-generation online connected systems, especially in the case of social networks, a relevant innovation of our times. This phenomenon is largely recognized as inducting relevant problems in modern societies, due to the pervasiveness of modern personal devices that originated larger and larger networks. Information and Communication Technologies (ICT) can really help to this end, thanks to the application of well-known models and methods mainly falling in the context of language transformers and neural networks, which both represent state-of-the-art solutions for supporting cyberbullying detection in text (e.g., posts). Inspired by this main research area, in this paper we provide an overview of state-of-the-art approaches for cyberbullying detection along with their experimental comparison against the reference TRAC-2 dataset.
Details
- Title
- Analysis and Experimental Comparison of State-Of-The-Art Deep-Learning Classification Techniques for Cyberbullying Detection
- Publication Details
- Procedia computer science, Vol.246, pp.3800-3809
- Resource Type
- Conference proceeding
- Conference
- International Conference on Knowledge Based and Intelligent information and Engineering Systems (KES 2024), 28th (2024)
- Publisher
- Elsevier B.V
- Number of pages
- 10
- Copyright
- © 2024 The Author(s).
- Identifiers
- 99381798267306600
- Academic Unit
- Center for Cybersecurity and AI; Hal Marcus College of Science and Engineering
- Language
- English