Preserving the Privacy of Sensitive Relationships in Graph Data Motivation Valuable Data! No privacy breaches! Public Data Anonymization Data representation? Privacy breach? Value of data? Data publisher Privacy Analyst Work by: Elena Zheleva, Lise Getoor, U of Maryland
Preserving the Privacy of Sensitive Relationships in Graph Data Data representation • Often, data is more complex , and it • Most approaches to data anonymization exhibits various relationships assume that the original data records are - E.g., friends, classmates, etc. in a single flat table Ana 21 F 20740 Bob 25 M 83201 Chris 24 M 20742 vs. Don 29 M 83209 Emma 28 F 83230 Fabio 31 M 83222 Gia 24 F 20640 Halle 29 F 83201 Ian 23 M 20760 John 24 M 20740 Possible privacy breaches • Identity disclosure • Sensitive relationship disclosure • Sensitive attribute disclosure - Sensitive relationship: one that has to remain private given the anonymized data Work by: Elena Zheleva, Lise Getoor, U of Maryland
Preserving the Privacy of Sensitive Relationships in Graph Data “Sensitive” link examples Search data Disease data Social network data Query 1: has hypertension “how to tell if your wife is cheating on you” father-of friends same-user Query 2: “myrtle beach golf course job listings” Anonymizing graph data • Goal : anonymize data (nodes and links) in a way to hide sensitive relationships • We propose 5 anonymization strategies (see our poster for details) • Compare the sensitive relationship disclosure in the anonymized data Link re-identification • Privacy breaches depend on the adversary domain knowledge • Assume the adversary can compute the probability of sensitive links • Probability is based on observed properties of the data Work by: Elena Zheleva, Lise Getoor, U of Maryland
Recommend
More recommend