Matrix Masking

Synonyms

Adding noise; Data perturbation; Recodings; Sampling; Synthetic data

Definition

Matrix Masking refers to a class of statistical disclosure limitation (SDL) methods used to protect confidentiality of statistical data, transforming an n × p (cases by variables) data matrix Z through pre- and post-multiplication and the possible addition of noise.

Key Points

Duncan and Pearson [3] and many others subsequently categorize the methodology used for SDL in terms of transformations of an n × p (cases by variables) data matrix Z of the form

$$Z\mathrel{\rightarrow }AZB + C,$$

((1))

where A is a matrix that operates on the n cases, B is a matrix that operates on the p variables, and C is a matrix that adds perturbations or noise.

Matrix masking includes a wide variety of standard approaches to SDL: (i) adding noise, i.e., the C in matrix masking transformation of equation [1]; (ii) releasing a subset of observations (delete rows from Z), i.e., sampling; (iii) cell suppression for...