Learning from Bullying Traces in Social Media

2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 656–666, Montr´eal, Canada, June 3-8, 2012. c⃝2012 Association for Computational Linguistics Learning from Bullying Traces in Social Media Jun-Ming Xu, Kwang-Sung Jun, Xiaojin Zhu Department of Computer Sciences University of Wisconsin-Madison Madison, WI 53706, USA {xujm,deltakam,jerryzhu}@cs.wisc.edu Amy Bellmore Department of Educational Psychology University of Wisconsin-Madison Madison, WI 53706, USA abellmore@wisc.edu Abstract We introduce the social study of bullying to the NLP community. Bullying, in both physi- cal and cyber worlds (the latter known as cy- berbullying), has been recognized as a seri- ous national health issue among adolescents. However, previous social studies of bully- ing are handicapped by data scarcity, while the few computational studies narrowly re- strict themselves to cyberbullying which ac- counts for only a small fraction of all bullying episodes. Our main contribution is to present evidence that social media, with appropriate natural language processing techniques, can be a valuable and abundant data source for the study of bullying in both worlds. We iden- tify several key problems in using such data sources and formulate them as NLP tasks, in- cluding text classification, role labeling, senti- ment analysis, and topic modeling. Since this is an introductory paper, we present baseline results on these tasks using off-the-shelf NLP solutions, and encourage the NLP community to contribute better models in the future. 1 Introduction to Bullying Bullying, also called peer victimization, has been recognized as a serious national health issue by the White House (The White House, 2011), the American Academy of Pediatrics (The American Academy of Pediatrics, 2009), and the American Psychological Association (American Psychological Association, 2004). One is being bullied or victim- ized when he or she is exposed repeatedly over time to negative actions on the part of others (Olweus, 1993). Far-reaching and insidious sequelae of bul- lying include intrapersonal problems (Juvonen and Graham, 2001; Jimerson, Swearer, and Espelage, 2010) and lethal school violence in the most extreme cases (Moore et al., 2003). Youth who experience peer victimization report more symptoms of depres- sion, anxiety, loneliness, and low self-worth com- pared to their nonvictimized counterparts (Bellmore et al., 2004; Biggs, Nelson, and Sampilo, 2010; Gra- ham, Bellmore, and Juvonen, 2007; Hawker and Boulton, 2000). Other research suggests that victim- ized youth have more physical complaints (Fekkes et al., 2006; Nishina and Juvonen, 2005; Gini and Pozzoli, 2009). Victimized youth are absent from school more often and get lower grades than nonvic- timized youth (Ladd, Kochenderfer, and Coleman, 1997; Schwartz et al., 2005; Juvonen and Gross, 2008). Bullying happens traditionally in the physical world and, recently, online as well; the latter is known as cyberbullying (Cassidy, Jackson, and Brown, 2009; Fredstrom, Adams, and Gilman, 2011; Wang, Iannotti, and Nansel, 2009; Vande- bosch and Cleemput, 2009). Bullying usually starts in primary school, peaks in middle school, and lasts well into high school and beyond (Nansel et al., 2001; Smith, Madsen, and Moody, 1999; Cook et al., 2010). Across a national sample …

Content truncated to 500 words