Cyberbullying is a type of bullying that uses technology such as cell phones to harass or malign another person. To detect acts of cyberbullying, we are developing an algorithm that will detect cyberbullying in SMS (text) messages. Over 80,000 text messages have been collected by software installed on cell phones carried by participants in our study. This paper describes the development of the algorithm to detect cyberbullying messages, using the cell phone data collected previously. The algorithm works by first separating the messages into conversations in an automated way. The algorithm then analyzes the conversations and scores the severity and frequency of the bullying words. A scoring threshold is used to predict whether or not a message or a conversation contains cyber bullying.
Bradley, Bryan W., "Detection of Cyberbullying in SMS Messaging" (2016). Computer Science Summer Fellows. 3.
Available to all.
Communication Technology and New Media Commons, Computational Linguistics Commons, Databases and Information Systems Commons, Social Media Commons, Theory and Algorithms Commons
Presented during the 18th Annual Summer Fellows Symposium, July 22, 2016 at Ursinus College.
Supported by a National Science Foundation Research at Undergraduate Institutions (NSF RUI) grant (No. 1421896).