Skip Header

Simulation Study of the Effectiveness of Masking Microdata with Mixtures of Multivariate Normal Distributions

Sam Hawala
Component ID: #ti1764678403


Continuous variables in microdata can be masked for protection from disclosure through the use of an additive noise. I consider adding noise that is distributed according to a mixture of normal distributions. There are several parameters involved in constructing the additive noise. The study’s purpose is to lay down as a guide a recipe for the choices of these parameters. The proportion of re-identifiable records through the use of Winkler’s matching software measures effectiveness of the masking method. Results depend heavily on the matching software used.

Related Information

  Is this page helpful?
Thumbs Up Image Yes    Thumbs Down Image No
Comments or suggestions?
No, thanks
255 characters remaining
Thank you for your feedback.
Comments or suggestions?
Back to Header