The goal of the one-year project is to determine how to test these systems and compare their effectiveness. The core of the project is the examination of data-mining tools.
"These are tools that look through different types of data and try to piece together a story," said Daniel Jeske, a principal investigator and UCR statistics professor. "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."
Because of the sensitivity of the project, the Department of Homeland Security has asked researchers not to describe in detail what kinds of data will be part of their project. But in general, data mining is the act of using statistics and computer technology to find patterns and connections in large amounts of gathered facts. The Department of Homeland Security can use such tools to better understand threat levels involving terrorism, and other national security interests.
The UC Riverside team also includes the statistics graduate students and staff of the newly formed Statistical Consulting Collaboratory, as well as James Lin, the Director of Computing of the UCR College of Humanities, Arts and Social Sciences. Jeske will work closely with Behrokh Samadi in the Performance Analysis Department of Bell Laboratories at Lucent Technologies.
"We had a kickoff meeting in Riverside," Jeske said. "We had someone from the Homeland Security side, Lucent sent a person and we had our team assembled. We currently have five students and four staff members being paid through the grant, using a combination of talent from the Department of Statistics and the Department of Computer Science."
Jeske, who came to UC Riverside in Fall 2003 after a 17-year career at Bell Laboratories, submitted the research proposal to the Department of Homeland Security while still at his former job. He brought the project with him when he came to UCR. Following a lengthy selection process, Homeland Security chose the joint collaboration of UC Riverside, Lucent Technologies and Bell Laboratories for the grant.
UC Riverside's slice of that grant is $320,000. The project ensures that UC Riverside students and staff will get hands-on, real-life experience in a project of significant magnitude. The research has applications in other realms besides security, Jeske said. "The tools could create a database on merchandise purchases, and establish a characterization of who is buying what," Jeske said. "Businesses can develop more efficient marketing strategies with that type of information." He said the project goal is to test the accuracy and validity of already developed data-mining tools. The project is challenging, Jeske says, because of the diversity in the data types used for information-discovery systems.
"The types of data vary greatly," Jeske said. "Our model can't be so specialized that it has to be completely redone for each individual use. It needs to be flexible and scaleable."
The presence of the Statistical Consulting Collaboratory, located in Sproul Hall, is vital to the project. The Collaboratory is a fee-based consulting service that assists with academic research, offers consulting services to the public and serves as a training center for graduate students.
"Having the Collaboratory definitely facilitates and enhances the ability to do this work," Jeske said. "This is a huge opportunity for our students to gain important experience in data mining. Having that on their resumes when they go to look for jobs makes them significantly more marketable. Another plus for the students is that they get to see how Bell Laboratories works, and work with some of the very best engineers in the world."
This is the first Homeland Security collaboration for the Statistics Department, and it will make it easier for the department to secure future government projects. "We're hoping for a renewal on this grant in 12 months," Jeske said. "We will build and implement so that the project extends itself. With this project, we are building a track record for the future."