基于B+树索引的动态社会网络差分隐私保护

针对当前社会网络的动态更新速度越来越快,而社会网络中差分隐私保护方法迭代速度慢的问题,提出一种基于B+树索引的动态社会网络差分隐私保护方法.使用B+树索引社会网络图的边,根据差分隐私并行性组合的特点,对B+树的索引数据划分,为数据分配不同的ε并添加拉普拉斯噪声,实现数据隐私后的整体高效用性和局部强保护性;在迭代时利用B+树的高效索引对欲更新的信息快速定位,实现动态社会网络差分隐私保护的快速迭代.实验表明,B+树索引有效提高了动态社会网络差分隐私保护的迭代速度,同时差分隐私的并行性提高了数据的效用性.

Dynamic social network differential privacy protection based on B+ tree index

The dynamic updating speed of social network is getting faster and faster. Aiming at the slow iteration speed of differential privacy protection methods in current social network, a differential privacy protection method of dynamic social network based on B+ tree index is proposed. By using the edge of B+ tree index social network graph, according to the characteristics of differential privacy parallelism combination, the index data of B+ tree are divided, and the data are assigned with different epsilon and epsilon ratios and added with Laplace noise, so as to achieve the overall high efficiency and strong local protection after data privacy. During the iteration, the efficient index of B+ tree is used to quickly locate the information to be updated, so as to realize the rapid iteration of differential privacy protection of dynamic social network. Experiments show that the B+ tree index effectively improves the iteration speed of differential privacy protection in dynamic social networks, and the parallelism of differential privacy improves the utility of data. This method effectively solves the problem of slow iteration of differential privacy protection in social networks.