Using Discussion Thread Context in Sentiment Analysis for Improving Cyberbullying Detection

Cyberbullying is the act of harassing, demeaning, or threatening someone via technology such as social networking, texting, or instant messaging. It is defined as “willful and repeated harm inflicted through the medium of electronic text” (Patchin and Hinduja 2006:152) [1]. Cyberbullying is an underlying problem in social networking services, threatening users' mental and physical health. The aim of this study is to improve the detection of cyberbullying in discussion threads by utilizing the repeated negative context formed in the presence of cyberbullying.