Cyber-aggression: user behavior detection, analysis, and prediction for multilingual Indian Social Media

Collaborator

• Dr. Rajesh Sharma, Associate Professor, University of Tartu
• Mr. Swapnil Mane, PhD Candidate, IIT Jodhpur

About

Cyberbullying is characterized as cyber-aggressive social media activity that includes harsh, rude, insulting, unpleasant, and sarcastic comments intended to harass others. This may be directed towards caste, religion, nationality, region, race, gender, color, etc. In many cases, targeted users suffer from many problems, such as low self-esteem, fear, anxiety, anger, suicide, and even physical attacks. A recent study by CRY (Child Rights and You) revealed that in India, eight out of ten individuals are victims of various forms of cyberbullying, of which $63%$ were abused and insulted online, while $59%$ were subjected to false rumors and gossip aimed at lowering their profile (CRY America Annual Report, CRY America, 2019). These findings are highly troubling as India being a land of a diverse population, may develop many complex problems out of cyber-aggression. The government provides sanctions under the IT Act and IPC 1850 (Sections 507, 509, 354(C&D), and 499) to monitor the adverse incidents by cyber-aggressive users on a day-to-day basis. Being persistently aggressive towards someone is a kind of user behavior (https://doi.org/10.1145/3343484). Many times people change their behavior because of their surroundings. For example, if someone sees people around them purchasing a specific brand of mobile, then they may also choose to buy the same. The question is whether the same is true for cyber aggression? In other words: Do my feeds make me aggressive? Further, it is also important to study whether aggression is based on the factor of segregation. That is to ask: Does aggressive behavior of users vary by language, region, gender, or any other distinct properties of the user or the target? Does my aggressive behavior depend on events (or topics)? Research to identify this behavior and change in behavior by digital means in the social age is a necessary and challenging task.

This research aims to propose an end-to-end solution for the exploration of potential cyber-aggressive behavior of users in an Indian multilingual context, which provides findings that will shed light on the novel psychological behavior of aggressive users on social media.