Data Sharing Taxonomy Records for Security Conservation

摘要

Here, we discuss the Classification is a fundamental problem in data analysis. Training aclassifier requires accessing a large collection of data. Releasing person-specific data, such ascustomer data or patient records, may pose a threat to an individual’s privacy. Even afterremoving explicit identifying information such as Name and SSN, it is still possible to linkreleased records back to their identities by matching some combination of non identifyingattributes such as {Sex,Zip,Birthdate}. A useful approach to combat such linking attacks, calledk-anonymization is anonymizing the linking attributes so that at least k released records matcheach value combination of the linking attributes. Our goal is to find a k-anonymization whichpreserves the classification structure. Experiments of real-life data show that the quality ofclassification can be preserved even for highly restrictive anonymity requirements.