Tag Archives: ICEWS

Steve Shellman is CEO and chief scientist at Strategic Analysis Enterprises, Inc., or SAE. In this role, Steve Shellman contracts with government agencies and corporations to provide statistical analysis, forecasting, and data collection services.

SAE helped to develop the Integrated Crisis Early Warning System, or ICEWS, as a tool to reliably predict political and social violence and upheaval across the globe. The development for ICEWS was contracted by the United States Defense Department, through the Pentagon’s Defense Advanced Research Project Agency (DARPA).

SAE contributed to three key parts of the data collection and analytic capabilities of ICEWS. It developed a tool called iTrace, which uses natural language processing algorithms to transform written content into quantitative data that reports interactions within and between countries. The iCast tool then uses data from iTrace and other sources to generate forecasts of actual events, such as crises, rebellions, and violence. The third component, iSent, is an opinion-mining tool that extracts information from web content about human responses to events or issues in a particular region.

SAE is currently working to extend ICEWS globally to create the Worldwide Integrated Crisis Warning System (W-ICEWS). During this multi-year project, SAE is expanding and updating the current system by improving the accuracy and transparency of forecasting models. New countries, sources, and languages will be available for analysis under W-ICEWS.

In 2008, Stephen M. Shellman founded Strategic Analysis Enterprises, Inc., to help support government agencies making national-scale decisions. The company has since worked on projects for the U.S. Department of Defense, the Joint Staff, and U.S. Strategic Command. It also did major subcontract work on the Integrated Crisis Early Warning System (ICEWS), a project funded by the Defense Advanced Research Projects Agency (DARPA).

DARPA funded ICEWS to create a platform that could reliably forecast upheaval and violence. It makes use of three major planks. First, it uses natural language processing to take unstructured news reports and turn them into quantitative indices. Second, it combines that data with other data to generate continuous forecasts of rebellions, international crises, insurgencies, and ethnic or religious violence. Third, it combs news feeds and blogs to measure public sentiment about events and organizations in a region.

ICEWS was created with a focus on the operations area of the U.S. Pacific Command. ONR extended the DARPA ICEWS project to the world. The Worldwide ICEWS or WICEWS project now includes global coverage of crises worldwide.