Samsung and Telefónica create algorithm to identify stalkers
Both companies developed a formula for Twitter, which allows to combat the actions of cyberbullying
A group of academic researchers, with the support of Samsung and Telefónica, developed an algorithm that is able to automatically identify Twitter accounts that carry out cyberbullying actions against other people, achieving an effectiveness of more than 90% in this task.
Experts recently published their study Detecting cyberbullying and cyber-aggression in social networks, in which they have analyzed 1.2 million Twitter accounts and 2.1 million tweets on topics such as the NBA or inequality of sexes in wages, around which abuse is more frequent.
Among these issues, the GamerGate also stands out, a harassment campaign that took place in 2014 against several developers, videogames and women in the sector who were targets of rape and death threats by a group of users on Twitter through the hashtag ‘#GamerGate. Five editors were banned from Wikipedia in 2015 for giving their page on this topic a feminist approach.
The study methodology is able to distinguish between aggressors and other Twitter users taking into consideration the text of their publications, their account and other network attributes, as detailed in the publication.
The study used the Twitter Streaming API, which provides access to 1% of all platform tweets. In addition, they filtered the spam accounts managed by bots and not by people, and the publications are organized in four types: stalkers, aggressive, spam and normal, which means without abuse.
K. Tovar
Source: dpa
