|
Activity Number:
|
513
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Government Statistics
|
| Abstract - #304169 |
|
Title:
|
Application of the Truncated Distributions and Copulas in Masking Data
|
|
Author(s):
|
Rahul A. Parsa*+ and Jay J. Kim and Myron Katzoff
|
|
Companies:
|
Drake University and National Center for Health Statistics and National Center for Health Statistics
|
|
Address:
|
CBPA, 2507 University Ave, Des Moines, IA, 50311,
|
|
Keywords:
|
truncated triangular distribution ; masking ; copula ; confidentiality
|
|
Abstract:
|
There are two main approaches to masking microdata---adding independent noise and multiplying by independent noise. The truncated triangular distribution, investigated by Kim (2007), was proposed for masking microdata with multiplicative noise: the masking variable is centered at one and truncated symmetrically around one so as to exclude values considered to be too close to one, thereby enhancing confidentiality protection. This paper generalizes Kim's earlier efforts by introducing truncated distributions of correlated noise in both the additive and multiplicative cases. Our method for adding correlated noise enables us to reproduce the first two moments of the original variable. We applied copulas to generate correlated noise. Two examples illustrate the methodology.
|
