An efficient hash-based algorithm for minimal k-anonymity

A number of organizations publish microdata for purposes such as public health and demographic research. Although attributes of microdata that clearly identify individuals, such as name and medical care card number, are generally removed, these databases can sometimes be joined with other public databases on attributes such as Zip code, Gender and Age to re-identify individuals who were supposed to remain anonymous. "Linking" attacks are made easier by the availability of other complementary databases over the Internet. k-anonymity is a technique that prevents "linking" attacks by generalizing and/or suppressing portions of the released microdata so that no individual can be uniquely distinguished from a group of size k. In this paper, we investigate a practical model of k-anonymity, called full-domain generalization. We examine the issue of computing minimal k-anonymous table based on the definition of minimality described by Samarati. We introduce the hash-based technique previously used in mining associate rules and present an efficient hash-based algorithm to find the minimal k-anonymous table, which improves the previous binary search algorithm first proposed by Samarati.