diffman: 'diffman' package
In diffman: Detect Differentiation Problems

Package to manage differentiation

Differentiation is a technique to deduce information on small agregates from a source of information that has been disseminated according to two different nomenclatures. Knowing the value of additive variables on agregates of observations according to the nomenclatures, one can decude the value of these variables on new agregates. If those agregates are smaller than a given threshold then we can considerer there is a confidentiality break for the observations within those agregates. It order to respect statistical secrecy for every observations, one need to check for the possible problematic differentiations that a data user can compute. The objective of this package is to provide a tool for detecting all observations at risk of differentiation.

diffman documentation built on Feb. 28, 2020, 5:08 p.m.