University of Colorado, Boulder (CU Boulder) researchers have developed a cyberbullying identification technique that blends multiple computing tools to scan massive volumes of social media data, alerting parents or network administrators of abuse.
CU Boulder's CyberSafety Research Center first had people teach a computer algorithm to differentiate between innocuous and abusive online comments, and then designed a system whose operation is similar to triage: when a user uploads a new social media post, tools quickly scan the comments and prioritize further checking if the comments appear questionable.
Real-time tests on data from the Vine video-sharing platform and Instagram found the system could flag cyberbullying behavior with 70% accuracy, and post alerts within two hours after such abuse is identified.
From CU Boulder Today
View Full Article
Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA
No entries found