Bully.tn: A Cyberbullying Detection Dataset in Tunisian Dialect
Fériel Ben Fraj Trabelsi, Sarra Kouki · 2024
The growing global use of social media has given rise to some problems, such as cyberbullying, which is a form of harassment intended to cause victims to constantly denigrate themselves, isolate themselves from others, or even take their own lives. Preventing the spread of this issue via social media is becoming more and more important. Gathering the necessary resources to comprehend and identify the many forms of dangerous communications is a prerequisite for intelligent detection. To detect cyberbullying in Tunisian dialect talks, this work presents a new dataset called Bully.tn. The set includes over 11,000 labeled comments from Facebook and YouTube. Using machine learning and deep learning techniques, it was discovered that our dataset was exceptionally semantically rich and relevant to the goal task.