Abstract

How to realize update in K-anonymity model

Author(s): Jinling Song, Liming Huang1, Gang Wang1, Qianying Cai1, Yan Gao

K-anonymity is a typical privacy model which can guarantee the safety of publishing dataset, however, the k-anonymized dataset contains generalized value and it difficult to bring it into correspondence with the original dataset directly. We at first create the index table basing on the one-one mapping between original tuple and its generalized tuple, which can be used to update the generalized tuple. To locate the QI group where an original tuple is in or should be inserted in, the definition of tuple-QG semantic similarity degree is presented and the QI group is located basing on tuple-QG semantic similarity degree. To merge the QI group whose size is smaller than k, QG semantic similarity degree are presented and used to find the similar QI group. Finally, the update algorithms basing on Semantic for the k-anonymized dataset are presented