Silva, Y.N., Hall, D.L. & Rich, C. Soc. Netw. Anal. Min
Springer International Publishing AG
2018-03-14
https://doi.org/10.1007/s13278-018-0496-z
Cyberbullying is the most common online risk for adolescents. While the prevalence and determinants of cyberbullying have received considerable attention among researchers in the psychology community, there has been relatively little work on the automatic identification of cyberbullying in social networking sites, and even less work that seeks to bridge the efforts from computer science and psychology. This paper thus proposes a computational model for cyberbullying identification that builds on the research findings within the psychology literature. The paper also describes the design of BullyBlocker, an app that implements the proposed model, discusses the model’s effectiveness in the context of a newly developed evaluative framework, and presents several ways in which the model can be extended. Our hope is that BullyBlocker, which has been recently made available through the Apple App Store, will have a strong societal impact, by identifying youth most vulnerable to cyberbullying victimization and by enabling parents to help their children in time to make a difference.