Enabling Accurate Analysis of Social Network Data while Protecting Privacy

04 Dec
Thursday, 12/04/2008 6:00am to 7:00am
Ph.D. Dissertation Proposal Defense

Michael Hay

Computer Science Building, Room 151

This thesis explores the tension between accurate analysis of social network data and the individual's right to privacy. While there has been a lot of research on privacy-protecting data mining, for the most part, this work assumes the data consists of a table of records, each corresponding to an independent entity. With social network data, there are relationships between the entities, and these relationships significantly change the problem, in terms of both the threats to privacy and the ways the data is analyzed. To address this challenge, we make the following contributions: an algorithm for assessing the privacy risks of data publication; a network anonymization algorithm that ensures node anonymity while preserving key topological features of the networks; and a query interface that allows a user to query a private network, returning accurate answers to a range of practical queries without disclosing the presence of individual relationships.

Dissertation Proposal Defense Required for Ph.D.
Advisors: David Jensen and Gerome Miklau