Magazine article Security Management

Time to Give Data Mining the Shaft?

Magazine article Security Management

Time to Give Data Mining the Shaft?

Article excerpt

DATA MINING and privacy. Some say the concepts are in irretrievable conflict. Just think back to the Total Information Awareness debacle and the continuing controversy over airline passenger screening via the use of government and commercial databases.


Proponents and opponents of the use of data mining for predicting terrorist events jousted recently on Capitol Hill before the Senate Judiciary Committee. Driving some of the debate was a policy analysis of predictive data mining published in December by Jeff Jonas and Jim Harper of the Cato Institute. The paper concluded that "Protecting America requires no predictive-data-mining technologies."

Harper supported this conclusion in testimony before the committee, noting that predictive data mining works when it finds a pattern that is consistent with "bad behavior" and inconsistent with "innocent behavior." He testified that "the terrorism context has a distinct lack of historical patterns" that link specific behaviors and preparation for terrorist acts.

Not only is data mining a blunderbuss tool, but even a highly accurate data-mining system would yield a tremendous number of false positives, Harper said. Data mining would thus "cost too much money, occupy too much investigator time, and do more to threaten civil liberties than is justified by any improvement in security it would bring."

Harper also dismissed red teaming--which he defined as the process by which security forces plan an attack and then identify data produced during the planning process. The pattern in the data might indicate preparation for that sort of attack. The problem with red teaming, he said, is that while it might identify terrorist planning, it will identify much more innocent planning, such as the activity of photo and architecture buffs who like to snap photos of tall buildings. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.