k-anonymity a model for protecting privacy.pptVIP

k-Anonymity: A Model for Protecting Privacy * Latanya Sweeney Carnegie Mellon University Intl. Journal on Uncertainty 2002 Presented by – Munawar Hafiz March 14, 2006 Information Security Internet and networking technologies have made the access to information much easier. Correlation of information compromises privacy. Information security is also related with confidentiality, integrity and availability. Information Security Integrity Confiden--tiality Availability User Preference Usability Proof of compliance Slide 1 of 19 Overview of the Presentation Re-identification of Data Terminology Introduction to k-Anonymity Attacks against k-Anonymity l-diversity Slide 2 of 19 Re-identification of Data Slide 3 of 19 Disease Birth Date Sex Zip Name 87% of the population in the USA can be uniquely identified by zip, sex and DoB. Terminology Tuple – A row of data Attribute – A column, A semantic category, A domain Inference – Belief on a new fact based on some other information Disclosure – Explicit and inferable information about a person. Disclosure Control – Attempt to identify and limit disclosures. Slide 4 of 19 Quasi-identifier – A minimal set of attributes in table that can be joined with external information to re-identify individual records. Terminology (continued) Frequency Set Slide 5 of 19 Select count(*) from patients group by sex, zipcode k-Anonymity k-Anonymity A relation is said to satisfy k-Anonymity property if every count in the frequency set is greater than or equal to k. The relation is called k-Anonymous. Slide 6 of 19 In plain English, a row in a table cannot be distinguished from at least k other rows. Generalization and Suppression Generalization A value is replaced by a less specific/more general value that is faithful to the original. Suppression Imposing on each value generalization hierarchy a new maximal element atop the old maximal element. Slide 7 of 19 Z0 = {53715. 53710, 5