Achieving k-anonymity using Minimum Spanning Tree based Partitioning
K. VenkataRamana, V. Valli Kumari · International Journal of Computer Applications · 2013
Protecting individual"s privacy has become a major concern among privacy research community.Many frameworks and privacy principles were proposed for protecting the privacy of the data that is being released to the public for mining purpose.k-anonymization was the most popular among the proposed techniques in which the sensitive association between the sensitive attributes and their corresponding identifiers are de-associated.In this paper, we proposed an enhanced k-anonymity technique by using Minimum Spanning Tree (MST) partitioning approach.In this technique we disclose the information of the individuals pertaining to minimum group size i.e., k.We achieve this technique in two phases.During the first phase, MST for the given dataset is partitioned to generate equivalence classes and in the subsequent phase whether the equivalence class size is achieved to that of the minimum group size k is verified.Our approach resulted in achieving the optimal anonymization along with data utility.We showed the efficacy of our proposed technique by running a series of experiments in terms of information loss to show that our technique adheres to the quality of the anonymized data.