|
Activity Number:
|
193
|
|
Type:
|
Invited
|
|
Date/Time:
|
Monday, August 3, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #303093 |
|
Title:
|
Differentially Private Categorical Data Analysis
|
|
Author(s):
|
Aleksandra B. Slavkovic*+ and Ivan Simeonov and Duy Vu
|
|
Companies:
|
Penn State University and Penn State University and Penn State University
|
|
Address:
|
Department of Statistics , University Park , PA, 16802,
|
|
Keywords:
|
Privacy ; Confidentiality ; Contingency Tables ; Disclosure Limitation ; Log-Linear Models
|
|
Abstract:
|
The concept of differential privacy as a rigorous definition of privacy has emerged from the cryptographic community. We evaluate how the proposed ideas of differential privacy can be applied to discrete response data such as Binomial and Poisson random variables. In particular we explore utility and validity of statistical analysis of contingency tables and log-linear models, and focus on sample size calculation such that we achieve both statistical efficiency (the confidence level and the power) and differential privacy. We consider advantages and disadvantage of the new proposed framework when compared to more traditional statistical disclosure limitation approaches when dealing with categorical data.
|
