US researchers accurately predict extremist behavior

SAN FRANCISCO (AA) – Scientists said Thursday they have created a computer framework that predicts global extremist attacks with more than 90 percent accuracy.

The creators of the framework examined more than 150,000 extremist attacks around the world that occurred between 1970 and 2015.

The Networked Pattern Recognition (NEPAR) framework catalogues data such as the time of day of an attack and weapons used. It detects patterns of behavior from the massive trove of data and paired with information like social media activity, it could help governments prevent future attacks.

By running NEPAR, its creators were able to predict most of the characteristics of attacks with more than 90 percent accuracy. By analyzing the entire network of data related to previous extremist attacks, NEPAR was able to identify the size of an attack with 90 percent accuracy and the goals of the extremists involved with 92 percent accuracy.

The project was led by Salih Tutun, a PhD student at Binghamton University in New York. A paper about NEPAR was published in the journal Expert Systems with Applications.

"Terrorists are learning, but they don't know they are learning,” Tutun said in a statement. “If we can't monitor them through social media or other technologies, we need to understand the patterns. Our framework works to define which metrics are important.”

Most of the research around predicting attacks focus on monitoring the activities of individual extremists and learning their patterns of behavior. NEPAR takes a more comprehensive approach by looking at historical data.

Armed with that data, law enforcement agencies can predict where an attack may happen or if a group will likely attempt multiple consecutive attacks.

"Predicting terrorist events is a dream, but protecting some area by using patterns is a reality,” Tutun said. “If you know the patterns, you can reduce the risks. It's not about predicting, it's about understanding,” he added.