基于并行聚类的物联网k-匿名算法

Parallel clustering-based k-anonymity algorithm in Internet of things

给出物联网可容忍空间粒度和可容忍时间粒度两个参数,建立适用于物联网环境的k-匿名模型。提出数据集分布序列的概念,以优化生成聚类种子,对数据进行并行聚类,使等价类含有多个节点的数据,具有布局特性的数据被划分到不同的等价类中,从而模糊数据的具体位置信息,消除标签的布局特性,设计适用于物联网环境的隐私保护k-匿名算法。实验结果表明,该算法在确保物联网数据可用的前提下,能够有效保护数据中的隐私信息,提高数据的安全性。

Two parameters about tolerable space granularity and tolerable time granularity in the Internet of things （IOT） are given and a k-anonymity model for IOT environment is created. P~ concept of the distribution sequence of data sets is proposed to optimize the generated cluster seeds. The data are clustered in parallel such that the data for multiple nodes are contained in the equivalence class, the data with specific location information will be divided into different equivalence classes to fuzzy their specific location information, the label layout characteristics is eliminated, and a k-anonymity algorithm for privacy protection in Internet of things is designed. The experimental results show that the presented algorithm can effectively protect the privacy of data and improve data security in premise of ensuring data availability in Internet of things.