基于进化算法的社会网络数据k-匿名发布

Evolutionary Algorithm for Social Network Data Publication Based on k-anonymity

基于科学发展的需要,越来越多的社会网络数据被共享发布。为保证发布数据中个体的隐私不被泄露,必须将数据进行隐私保护后发布。针对结点度的再识别攻击,提出一种改进的进化算法对社会网络发布的数据进行k-度匿名(CEAGA),将EAGA算法中的适应度函数与循环结束条件进行改进,得到最优的k-度匿名序列,之后按照得到的k-度匿名序列对匿名图进行构造,得到最优的k-度匿名社会网络图。实验结果表明,改进后的进化算法不但降低了对原社会网络图的修改,并且对图结构性质的保持也优于EAGA算法。

Based on the needs of scientific development, a growing number of social network data to be shared and released. In order to ensure that the privacy of the individual ＂s privacy is not leaked, privacy protection should be carried on before releasing social networks data. For re-identification attack of the node degree, we pro- posed an improved evolutionary algorithm that to carry out the k-degree anonymity for social networks data. This paper improved fitness function and end condition of the loop of EAGA algorithm , and obtained optimal k-degree anonymous sequence. Then we obtained the optimal anonymous social network graph of k-degree by construct a- nonymous graph based on k-degree anonymous sequence that previous algorithms was generated. Experimental re- suits show that the improved evolutionary algorithm not only reduces the modification of the original social network graph, keep the property of the graph structure is better than EAGA algorithms.