Privacy Preservation Based on Key Attribute and Structure Generalization of Social Network for Medical Data Publication

摘要

Protection of privacy data published in health care field is the research to prevent the problem of disclosing sensitive data in healthcare. The Health Insurance Portability and Accountability Act (HIPPA) in the USA is the current practice regulation of privacy protection. Recently, however, the Institute of Medicine Committee on Health Research and the Privacy of Health Information concluded that HIPPA cannot adequately safeguard privacy and allow researchers to effectively use them for discoveries at the same time. Privacy protection method based on clustering is the process of developing methods and algorithm to ensure that the published data remains useful and protected. In this paper, we purposed an algorithm based on greedy clustering to group the data points according to the attributes and the connective information of the nodes in the published social network. During the procedure of clustering, we handled the loss of information, and evaluated the proposed approach in terms of classification accuracy and information loss rates on real medical datasets. The experimental results in our proposed approach show that the data privacy can be protected with less information loss. Finally we show a visualization process for the clustering.