New algorithm could predict ISIS attacks

PanARMENIAN.Net - While much of the war against terror takes place in the air and on the ground, a new tool for fighting extremists does all its work online, Digital Spy said.

A physicist from the University of Miami has created an algorithm that can scan the Islamic State's social media posts, a function that may eventually be used to predict the group's next terrorist attack.

In the journal Science, Dr Neil Johnson and his team described the method of searching through roughly a year's worth of posts on the Russian-based social network Vkontakte, looking for pro-ISIS statements in different languages.

The resulting data allowed Johnson and team to create a statistical model aimed at identifying behavioural patterns among online supporters of ISIS.

Rather than trawling through an ocean of social media posts looking for nuggets of information about a potential attack, though, Dr Johnson's model looks at how small, self-organised groups appear online in advance of real-world campaigns.

Per the New York Times, these groups "reflect groundswells of new activity and, if followed, can potentially point to where that activity is going".

While it would be unrealistic for the model to predict incidents like San Bernardino or Orlando, where one or two people acted alone, it could be used to stop groundswells before they gain enough momentum to become a more widely influenced terror attack.