2.14 Security and Privacy

A major concern is the possible misuse of all of this information and the need to ensure that the models that evolve from it are correctly applied and are rigorously tested before being implemented. Any profiles require that ample tests be performed prior to widespread deployment. The American Civil Liberties Union (ACLU), for example, is raising concerns about plans for the Transportation Safety Administration's CAPP II (computer-assisted passenger pre-screening) system to pull in data from banks, and other data providers mentioned in this chapter. Care must be taken to ensure the security and privacy of personal data is protected and the rights of citizens are not violated.

Issues involving security, especially when it comes to Internet, financial, credit, and government data, will also influence any data-integration efforts. Copyrights, personal and medical records, and other sensitive data will likely also restrict what information can be used. Above all, the privacy and security of individuals must be protected, and any access to private information, such as unlisted telephone numbers, must be done with the proper legal instruments, such as subpoenas or summonses. Note that some data mining models can be constructed without specific identification information, such as names, addresses or other unique identification numbers. To further protect the data, it can be classified using a variety of methods from unrestricted to a special compartmentalization-restricted security level.