Communications of the ACM
Researchers Can Predict Terrorist Behaviors With More Than 90 Percent Accuracy
Researchers at Binghamton University developed a framework to predict future terrorist behaviors by defining past attack patterns.
Credit: oneinchpunch/Shutterstock
Researchers at Binghamton University developed the Networked Pattern Recognition (NEPAR) Framework to predict future terrorist behaviors by defining past attack patterns.
The team trained NEPAR on more than 150,000 terrorist attacks between 1970 and 2015 so it could calculate relationships among attacks and spot distinctive behaviors with these links.
The framework first constructs networks by finding connections between events, and then employs a unified detection strategy integrating proposed network topology and pattern-recognition approaches.
NEPAR identifies the traits of future terrorist attacks by analyzing how past attacks relate, and a comparison of its results with existing data found the framework could predict most attack characteristics with more than 90% accuracy.
In addition, the researchers say their unified detection approach could be used to apply pattern classification methods to network topology and features of incidents to spot terrorism attacks with high accuracy, and identify the extension of attacks, multiple attacks, and terrorist objectives.
From Binghamton University
View Full Article
Abstracts Copyright © 2017 Information Inc., Bethesda, Maryland, USA
No entries found