Dataset could be divided into several equivalent sets according to K-anonymity, and each set contains at least K and no more than 2K records.

In unifying phase, we check all equivalent sets in the dataset at first and then check each tuple in the equivalent set.

The clustering K-anonymity would assign similar records to the same equivalent set, while the similarity among these records makes it harder to discriminate different identities than before.

(3) Link {[QI.sub.1], [QI.sub.2], ..., [QI.sub.n], SD} and |ID, [QI.sub.1], [QI.sub.2], ..., [QI.sub.n]} to find the equivalent set ES which contains the ID of objective.

In our work, we try to adjust the division of records, making it hard to discriminate the identity within each equivalent set, thus preserving privacy.

In this paper, for preserving users' privacy, we expect the records in the same equivalent set to be as similar as possible.

At first, we cluster the records in private table which need to be published, and assign similar records to the same equivalent set. Then, we unify the quasi-identifiers in the same clusters by generalizing and suppressing operations.

The goal of this method is to preserve privacy, so a lower discriminating rate within one equivalent set suggests a higher security performance.

For example, a few days prior to the lesson, Kathleen had asked the children to choose any number between 10 and 30 and find as many ways as possible to partition that number into

equivalent sets. In the whole group discussion, she began with, "As I went around I noticed many of you chose 20 and 30.