Towards Automatic Cybertroll Detection in Social Media
Rama Seetha Maha Lakshmi Patibandla · 2024
Cybertrolling is a deliberate and repeated act of aggression through various social networking sites such as Instagram, Twitter, Facebook, etc. Bidirectional Encoder Representation (BERT) is a cutting-edge pre-trained approach that performs excellently in a number of languages empathetic errands. In this chapter, we propose a new BERT application for cybertrolling detection. BERT is a deep learning model used to achieve the best results diagonally in three real-world datasets like Twitter, Formspring and Wikipedia. Experimental outcomes show that our planned representation outperforms existing works, compared to the attention-based or slot-gated deep neural network models.