A new machine learning algorithms have been developed by team of researchers, including faculty at Binghamton University which can successfully identify bullies and aggressors on Twitter with 90 percent accuracy. The effective tools for detecting harmful actions are rare in social media as this type of behavior is often ambiguous and/or exhibited via seemingly superficial comments and criticisms. Research team analyzed the behavioral patterns exhibited by abusive Twitter users from other Twitter users differences aimed at bridging this gap. Researchers collect data with crawlers - programs built from Twitter via variety of mechanisms, gathering tweets of Twitter users, as well as social network-related things, like who they follow and who follows them. After this, natural language processing and sentiment analysis along with a variety of social network analyses performed on the connections between users. Two specific types of offensive online behavior – cyber bullying and cyber aggression are classified by developed algorithm, to identifying abusive users on Twitter with 90 percent accuracy. The damage being done to humans is one of the biggest issues with cyber safety problems and is very difficult to undo. At present, the team are currently exploring pro-active mitigation techniques to deal with harassment campaigns.
0 Comments