Statistical Disclosure Control for Tabular Data in R

Kazuhiro Minami (
The Institute of Statistical Mathematics / National Statistics Center, Japan
Yutaka Abe (
National Statistics Center, Japan


To perform statistical disclosure control (SDC) on tabular data is a challenging task because we need to ensure that every suppressed cell of a table has a sufficient width of a confidentiality interval under the presence of linear relations among cell variables. However, we find that the existing SDC tool (i.e., τ-ARGUS) does not effectively support an output checking process of the on-site use program in Japan. We therefore develop a new SDC tool in R, which produces safe tabular data with auxiliary information that is necessary for an output checker to verify its safety.
In this paper, we describe the major features of our SDC tool and discuss possible extensions in the future. Our SDC tool performs primary suppressions on a frequency table and a magnitude table with the minimum frequency rule and an occupancy rule (e.g., (n,k)-rule), respectively. We implement the optimal secondary suppression mechanism based on the technique of Benders decomposition.

Keywords: Statistical Disclosure Control, Output checking, Cell suppressions, Linear programming
JEL Classification: Z00

[Full Text]

Romanian Statistical Review 4/2017