Data scientists have designed an early-warning model that can successfully foretell how deadly a terror organization will become in the future founded on only its first ten attacks.
Researchers from Northwestern University’s Kellogg School of Management developed the predictive model which will allow security forces to better detect and aim the most hostile groups to possibly stop them before they grow too influential.
“This early forewarning is huge because not only can it help the government aim and offset the groups with the most potential for destruction, it also can aid the government tactically use resources and avoid wasting billions of dollars fighting a group that is likely to exhaust on its own anyway,” said Brian Uzzi, corresponding writer on the study and the Richard L. Thomas Lecturer of Leadership and Organizational Change at Kellogg.
The study named “Quantifying the Future Lethality of Terror Organizations,” was published on Oct. 7 in PNAS.
The concept, which uses information publicly available through the RAND Database of Worldwide Terrorism Incidents (RDWTI) and the Global Terror Database (GTD) and, has substantial power in forecasting a terror organization’s lifetime ferocity after just ten terror assaults.
Researchers adjusted the model using facts from terror groups functional between the years of 1970 and 2014. They discovered that some of the most thought-provoking model predictions were for groups that went with very few attacks in the beginning, only becoming fatal much later. Among these groups were Al-Shabaab, United Liberation Front of Assam and Moro Islamic Liberation Front.
“The model can predict the future influence of some of these sleeper groups even while they are still functioning covertly,” said research co-lead author Yang Yang, a postdoctoral associate at Northwestern.
The United States government employs half a trillion dollars yearly to research and fight terrorism. According to information from GTD, between the years of 2000 and 2015, 61 new terror groups arose each year, on average, leading to an 800% rise in global terror attacks.
“Former models, for the most part, are useful for comprehending the situation in which terror activity is expected to take place, but they are too limited to the locale and not beneficial in predicting individual organizations’ conduct,” said Adam Pah, co-lead author of the research paper and clinical associate professor of management and organizations at Kellogg.
The researchers turned to the business world for motivation for a better model.
“Basically we said, ‘What if we think of terror groups like a business whose product is lethality? How do we forecast their success in producing that product?'” Uzzi said.
Business investors and venture capitalists regularly use publicly reported information such as technological skills and cash flows to predict company success and behavior. This information is not present for private terror organizations, so the researchers worked to develop substitutes based on visible behavior.
For instance, business investors frequently view the timing of a company’s product launches as a proxy for resources. They presume that a company that recurrently launches new products likely has more means than a company that brings new products at random. Likewise, the researchers’ model uses the timing of attacks as a representation of a terror group’s organizational strength and resources.
The researchers were able to confirm these concepts with elements like the range of weapons used, the sophistication of those weapons, and their attack abilities, defined as the magnitude to which the group was successful in executing the mission of the attack.
Researchers found that Islamic State had astonishingly strong attack capabilities near the 90th percentile of all terror groups in similar years, even though the group proved an irregular attack rhythm that was suggestive of unstable resources. After just 10 attacks, the model positioned ISIS among the terror groups with the most potential for committing extremely deadly attacks.