Sign In

Communications of the ACM

ACM TechNews

Algorithm Can Distinguish Cyberbullies from Normal Twitter Users with 90% Accuracy

View as: Print Mobile App Share:
Cartoon of a cyberbully.

Researchers at Binghamton University have developed machine learning algorithms that can identify bullies and aggressors on Twitter with 90% accuracy.


Machine learning algorithms developed by a team of Binghamton University researchers can differentiate bullies and aggressors on Twitter from normal users with 90% accuracy.

The researchers analyzed behavioral patterns of abusive Twitter users and how they differ from the activities of other Twitter users.

Binghamton's Jeremy Blackburn said the researchers then used crawlers, “programs that accumulate data from Twitter via a variety of mechanisms,” to gather the tweets and Twitter profiles of users, “as well as social network-related things, like who they follow and who follows them.”

The application of natural language processing and sentiment analysis on the tweets, along with social-network analyses of user connections, led to algorithms that automatically classified cyberbullying and cyberaggression.

Blackburn said, "In a nutshell, the algorithms 'learn' how to tell the difference between bullies and typical users by weighing certain features as they are shown more examples."

From Binghamton University News
View Full Article


Abstracts Copyright © 2019 SmithBucklin, Washington, DC, USA


No entries found