机构地区: 广州大学数学与信息科学学院
出 处: 《计算机应用与软件》 2009年第12期114-116,142,共4页
摘 要: 基于泛化映射的k-匿名技术是保护数据共享环境中的隐私信息的有效方法。然而,寻求具有最佳可用性的泛化映射是NP困难的。提出了一种泛化映射的局部搜索算法,从满足k-匿名的一个等深泛化映射向量出发,依据预定的属性次序,寻找并分解选定属性的过泛化映射,形成新的非等深的满足k-匿名的泛化映射向量。实验表明在搜索等深映射的时间开销基础上,增加少量的计算时间,就可得到可用性更高的泛化映射。 The k-Anonymity technique, based on generalization mapping, is an efficient method to preserve information privacy in public data sharing environment. However, to find a generalization mapping with the best utilities is of NP hard. This paper proposes a local search algorithm of generalization mapping, proceeding from a depth-equal generalization mapping magnitude satisfying the k-anonymity to find and decompose the over-generalization mapping of the selected attributes according to a predefined attribute order, and forms new non depth-equal generalization mapping vector satisfying the k-anonymity. The conducted experiments show that the generalization mapping with higher utilities can be found by only increasing limited computation costs based on the time cost of searching depth-equal mapping.