ABSTRACT
So far, we have considered the simplest case where the data is published to a single recipient. In practice, the data is often published to multiple recipients and different data recipients may be interested in different attributes. Suppose there is a person-specific data table T (Job, Sex,Age,Race,Disease, Salary). A data recipient (for example, a pharmaceutical company) is interested in classification modeling the target attribute Disease with attributes {Job, Sex,Age}. Another data recipient (such as a social service department) is interested in clustering analysis on {Job,Age,Race}.
