Abstract:

 The protection of sensitive knowledge contained in the data is a key research issue in privacy preserving data mining. Granular computing has a powerful ability in using different levels of granulations during problem solving. By using granular computing model derived by rough set, a method for the knowledge hiding in the data was discussed. First, some concepts such as tolerance relations, tolerance information granules and the size of granules were introduced. Second, an approach for knowledge hiding was proposed. Final, the given example and testing in real datasets showed that the proposed method was effective.
 

Key words: knowledge hiding; privacy protection; granular computing; tolerance information granule