What is it about?

Multiple sensitive labels without correlation is a simple scenario in which the data publisher can use any conventional privacy model which supports static data. Our work (SNI) anonymizes a social network having multiple sensitive labels with correlation. Proposed algorithm of SNI minimizes the membership error rate for membership analysis. SNI has been evaluated on cancer label of individual in an Iranian hospital.

Featured Image

Why is it important?

SNI publishes all labels of vertices without distorting almost with preserving correlation. It calibrates privacy and utility of data, by expert’s desired attack threshold.


Writing this article was a great pleasure, becuase preserving the correlation among labels and privacy is an important issue in publishing social networks which have multiple sensitive labels with correlation. Since, in real life, there is a correlation among multiple sensitive labels, for example, in medical data, the sensitive labels like disease, physician, symptoms, treatment are correlated and breaching one can easily breach the other labels. This article also lead to groups contacting me and ultimately to a greater involvement in supplying data utility.

privacy preserving on social networks Abbas Karimi Rizi
Islamic Azad University

Read the Original

This page is a summary of: SNI: Supervised Anonymization Technique to Publish Social Networks Having Multiple Sensitive Labels, Security and Communication Networks, November 2019, Hindawi Publishing Corporation,
DOI: 10.1155/2019/8171263.
You can read the full text:

Open access logo


The following have contributed to this page