The rise of online social networks (OSNs) has resulted in a significant increase in cyberbullying activities, especially among teenagers. Indeed, cyberbullying has become a major societal problem. Cyberbullying is such a serious problem that nine suicides have been linked with cyberbullying on the Web site alone. There is an urgent need for detecting incidents of cyberbullying in OSNs, so that appropriate help can be provided to the cyberbullying victims in a timely manner and bullies can be identified and stopped.


  An interaction graph can be constructed from the “likes” of answered questions from collected profiles. 

 A key finding in our research is that users labeled with high degree of negativity are distinguished by having the smallest total in-degree and the smallest total out-degree. As the amount of positive support increases, we find a greater in-degree and greater out-degree of the users, that is users become more social and actively like and are liked more often.

 We observed different types of cyberbullying:


 Among users, one of the most disturbing behaviors that we encountered  was the problem of “cutting” (slicing one’s wrists). Looking at the profiles of these cases and studying their answers, it seemed that the profile owner exhibited weak confidence and depression problems, sometimes admitting to earlier failed attempts at suicide.
Subpages (1): Dataset