During the process of data publishing, in order to protect personal privacy, data owner often has to generalize all the quasi-identifier, but in fact the tuple involving personal privacy is very few. So our method starts from the tuples, then divides the remainder into groups according to different sensitive attribute value. Finally, our method selects one tuple per group which is shortest to the tuple involving personal privacy for generalization. The rest tuples need not to be generalized. In this way, our method has improved the utilization rate of the data and reduced the loss of information effectively.%在数据发布的过程中，为了保护个人隐私常需对所有准标识符进行泛化操作，而实际涉及到个人隐私相关敏感属性元组是非常少的。据此，从这些涉及个人隐私的敏感属性的元组出发，将剩余大量仅涉及非敏感属性元组依据敏感属性值不同进行分组，最后对分组中元组以计算与个人隐私属性相关敏感属性距离的方式，选取距离最短的元组进行泛化，其余元组并不进行泛化，通过这种方式，提高了数据的利用率，并有效减少信息的损失。
展开▼
