Researchers Test Accuracy Of Homeland Security Data Mining Programs
The University of California, Riverside (UCR) is working with Lucent’s Bell Laboratories on a data-mining validation project that will help the Department of Homeland Security test the accuracy of behavioral patterning.
The one-year project is funded with an $800,000 grant from the Department of Homeland Security and will develop a model for testing how accurate data-mining programs are, especially data-mining used to identify patterns of behavior.
The U.S. government plans to use such technology to better understand potential national security threats, says UCR statistics professor and principal investigator for the project Daniel Jeske; “These are the tools that look through different types of data and try to piece together a story,” he says about the Department of Homeland Security data-mining tools that will be examined. “The tools say an event could happen based on patterns that are found in the data. Sometimes the tools are referred to as information-discovery systems.”
The data-mining accuracy model needs to be flexible to accommodate the widely varying types of data that will be used to identify behavioral patterns, and could be applied to commercial purposes in the future; marketers could use the system to more accurately characterize customer behavior, for example. The project will involve five graduate students and four staff members from UCR’s statistics and computer science departments, as well as researchers from Bell Laboratories.
Abstracted by the National Law Enforcement and Corrections Technology Center(NLECTC) from the UCR News (02/03/05).
