Abstract:

A great deal of information exists in kanonymous data. How to get useful information from kanonymous data is an urgent pending problem. OLAP (OnLine Analytical Processing) is the main approach of knowledge discovery, and the aggregate query is the key operation of OLAP. In order to solve the problem of aggregate query over kanonymous data, firstly, the definition of data model describing kanonymous data is given. Secondly, the aggregate query is separated into two phases. On the first phase, the properties of kanonymous data satisfication and the notion of Independent Attribute Set is presented. Using these properties and the Independent Attribute Set, an algorithm is given to compute the set of value and its probability that satisfy the query constraint, and then take the set as the input of second phase. On the second phase, the semantics of the aggregate query over kanonymous data are defined. In order to meet user’s different query, the definition and the semantic of WITH clause constraint is given as a supplement to first phase. At last, properties of the aggregate query are shown and an experiment is done to prove the validity of our method.

Key words: data sharing;on-line analytical processing;privacy preserve;kanonymity;aggregate query