Activity Number:
|
306
|
Type:
|
Topic Contributed
|
Date/Time:
|
Wednesday, August 14, 2002 : 10:30 AM to 12:20 PM
|
Sponsor:
|
Section on Government Statistics*
|
Abstract - #300536 |
Title:
|
Safely Disseminating Tabular Data: Table Servers and Optimal Tabular Releases
|
Author(s):
|
Alan Karr*+
|
Affiliation(s):
|
National Institute of Statistical Sciences
|
Address:
|
19 T.W. Alexander Drive, PO Box 14006, Research Triangle Park, North Carolina, 27709-4006, USA
|
Keywords:
|
data confidentiality ; statistical disclosure limitation ; digital government
|
Abstract:
|
Two classes of systems being developed by NISS for disclosure-limited dissemination of tabular data will be described. Table servers are Web services that respond to user queries for marginal sub-tables of a large contingency table, and feature dynamic evaluation of disclosure risk in light of previously answered queries. Issues include risk criteria, release rules and scalability. Optimal tabular releases (OTRs) are fixed sets of released marginal sub-tables constructed by maximizing a measure of data utility (of the released information) subject to a constraint on risk. Issues for OTRs are risk criteria and thresholds, measures of data utility and scalability. One approach to the latter is to require that the released marginals constitute the minimal sufficient statistics of a decomposable graphical model.
|