2020/09/30 by Mohit Chandra, Chandra, Mohit, Ashwin Pathak +11
Computer Science · #Computation and Language (cs.CL) #Cybercrime and Law Enforcement Studies #FOS: Computer and information sciences #Hate Speech and Cyberbullying Detection #Information Retrieval (cs.IR) #Spam and Phishing Detection
paper · pdf · doi:10.48550/arxiv.2010.00038
openalex publication_date 2020/09/30 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28
While extensive popularity of online social media platforms has made\ninformation dissemination faster, it has also resulted in widespread online\nabuse of different types like hate speech, offensive language, sexist and\nracist opinions, etc. Detection and curtailment of such abusive content is\ncritical for avoiding its psychological impact on victim communities, and\nthereby preventing hate crimes. Previous works have focused on classifying user\nposts into various forms of abusive behavior. But there has hardly been any\nfocus on estimating the severity of abuse and the target. In this paper, we\npresent a first of the kind dataset with 7601 posts from Gab which looks at\nonline abuse from the perspective of presence of abuse, severity and target of\nabusive behavior. We also propose a system to address these tasks, obtaining an\naccuracy of ~80% for abuse presence, ~82% for abuse target prediction, and ~65%\nfor abuse severity prediction.\n