The Research On Data Update Of K-Anonimity Table

K-anonymity is an important approach to prevent data from disclosing in view publishing. The view is named as k-anonymity table after it is anonymized. The basic table is always updated,so how to keep the data in k-anonymity table to be in accordance with those in the basic table is a valuable researched subject. Based on this, the problem about updates of the k-anonymity table is put forward, which aimed at improving the k-anonymized efficiency with method that the new k-anonymity table is gotten in a short period at the base of the previous version. This paper presents a comprehensive analysis based on the current situation of the domestic and international research about the technology in k-anonymity privacy protection and the problem about updates of the k-anonymity table needs further researches. The detailed content can be illustrated as follows.Firstly, this paper analyses several information attack approaches and information disclosing models, such as multi-party collusion and information inference. Three stages are introduced to guarantee the security of views publication. Based on the analysis in the drawback that the current k-anonymity table presents, the formulae of the privacy disclosing measure and the information preserving degree are given. An improved algorithm that takes the maximum privacy disclosing rate demand of the view publisher into consideration is provided.Secondly, according to the response that k-anonymity table makes when the base table of the database is updated, the updating operations of the k-anonymity table can be divided into three conditions: irrelevant update, autonomously computable update and routine update. The correlations between the updating operations and the k-anonymity table are researched detailedly when each updating operation is satisfied. Meanwhile the corresponding theory analysis, theorem proof and example explanation are introduced.Moreover, an incremental updating approach for k-anonymity table based on the dynamic tree is presented. This approach uses the dynamic tree as the basic structure, maps the published table to a union consisting of multi-dimensional space dots and translates the updating operation (insertion, deletion and modification) of the published table to the corresponding one of the dynamic tree. And an incremental updating approach for k-anonymity table based on the multi-dimension bucket is proposed. This approach analyses the changes that are brought by the updated tuples of the basic table to the multi-dimension bucket and maps the changes to the incremental updates of the k-anonymity table.Finally, extensive experiments are conducted to testify these proposed algorithms and the experimental result is given and analyzed carefully.