Applicability of Basic Separability Principles to Enhance the Operational Efficiency of Synthetic Tabular Data Generation Procedures in Multidimensional Table Structures

• Activity Number: 338
• Type: Topic Contributed
• Date/Time: Tuesday, August 2, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Committee on Privacy and Confidentiality
• Author(s): Ramesh Atmaram Dandekar*+
• Companies: Department of Energy, EI-20
• Keywords: Statistical Disclosure ; perturbation ; tabular data ; establishment data ; individual data ; privacy

Abstract

Hyperlinked copy of this abstract is available at URL http://mysite.verizon.net/vze7w8vk/AbstractJSM2011.pdf

Dandekar2001 proposes using synthetic tabular data generation (i.e. controlled adjustments to tabular data - CTA) as an alternative to complementary cell suppression procedures. Dandekar2009 addresses quality aspects of CTA protected tabular data with an objective to completely replace conventional complementary cell suppression procedures with a new tabular data protection method. The proposed method is a hybrid of three different tabular data protection methods. CTA has already been demonstrated to be equally effective on multi-dimensional counts data and magnitude data containing complex hierarchies and linked table structures.

In this paper we go a step further to demonstrate how the operational efficiency of the CTA procedures could be enhanced significantly by applying basic separability principles to complex table structures. The proposed enhancements have significant potential to cut down on computational resources by reducing the problem size in terms of the number of variables and associated mathematical constraints.