Abstract: Nowadays many kinds of personalized privacy protection algorithms are proposed. They are mostly for static data, but the new set and the anonymous set together dynamic update is a problem to be solved. It establishes a dynamic update method on [(α,k)]-anonymous data set. For each tuple it calculates the semantic and chooses the appropriate equivalence class to do tuple update; it makes final dynamically updated set satisfy[(α,k)]-anonymous. Algorithm analysis and simulation results show that this algorithm obtains dynamically updated implementation of smaller loss of information and short executing time.

Key words: data publication, privacy protection, [(α,k)]-anonymous, semantic closeness

摘要： 现如今已提出了多种个性化的隐私保护算法，这些隐私保护算法大多是针对静态数据的，而动态新增数据集和原始匿名数据集的同步更新是一个亟待解决的问题。建立一种在[(α,k)]-匿名数据基础上的增量更新方法，对于每个元组计算语义贴近度并选择合适的等价类进行元组更新，使得最终动态更新后的数据也满足[(α,k)]-匿名。算法分析及仿真实验结果表明，算法以较小的信息损失和执行时进行数据动态更新。

关键词: 数据发布, 隐私保护, [(&alpha, k)]-匿名, 语义贴近度