The k-anonymity approach for data-publishing based on clustering partition

k-anonymity is a widely used technology for privacy protection. In order to prevent attackers from digging out private information based on data published and his(her) background knowledge, a k-anonymity method is proposed, which uses the idea of clustering partition. Two different types of quasi-identifier attributes are individually subjected to a generalization process. The distance between tuples as well as the distance between a tuple and an equivalence class is also defined. In the process of clustering partition, some tuples are selected to constitute equivalence classes one by one, according to the principle of minimum distance value. k-anonymization is then completed, finally, by a separate generalization process. Through experiments comparing the performance of our method with that of an existing anonymization algorithm, EBKC, the effectiveness of the approach is verified.