Researchers at the University of Salamanca (USAL) have developed sentiment analysis algorithms that allow monitoring and locating violent groups on Twitter by analyzing the sentiment of the messages they share."This system could have been very useful as a support system to control-for example-the groups of violent fans that caused serious incidents during the celebration of the last Eurocopa in France, "Juan Manuel Corchado comments to Sinc , Professor of Computer Science and leader of the artificial intelligence team Bisite of the USAL.
According to Corchado, the police and other security forces can use the tool to detect critical points, threats and concentrations of potentially dangerous people."It is based both on the semantic analysis of the messages, as on the historical data and their evolution ".
For its development, the team has used artificial intelligence neurosymbolic hybrid systems and methods that solve problems using a mixture of different algorithms, "with good results," he emphasizes.
Analysis in different languages
The advantage of the application, says Corchado, is that you can do a sentiment analysis in different languages -for the moment, Spanish, English, French, German, Russian and Arabic-and, in addition, extract information about communications continuously and records how both the feeling and the location of the individuals evolve, while analyzing the interrelationships of these groups.
The system-anade-" is able to establish quite accurately where a dangerous user is located from what you share on Twitter and how and with whom you interact at all times, without the need to geolocate the tweets ".
Depending on the traffic generated between people, and s it is possible to identify the members of a group, the leaders and the followers.Also see how the relationships evolve and if new members join the group. In addition, once the leaders have been identified, you can try to influence them to modify their feelings, says the researcher.
Corchado ensures that the tool is in a very advanced stage."We have developed a prototype to experiment with new algorithms and functionalities.And we have almost ready the tracking system of individuals and their contact networks, capable of monitoring the evolution and the interaction between different people or groups.All this, in a very dynamic and easy-to-use format, "he emphasizes.
Prevention of bullying and racism
Other uses of the system could be the prevention of phenomena such as bullying and racism.It also serves to analyze what is said in the networks about companies, brands and people. Researchers at The USAL is already working with the State security forces, which will soon begin to validate the tool.."For now, I cannot move forward any more," says the professor.
Source: SINC Agency
Comments
Post a Comment