Description

The aim of statistical disclosure control is to keep up the required statistical privacy while making data available to the researchers. This can be achieved with the help of minimal modifications of the data without changing the multivariate data structure. In this book the well-developed R package sdcMicro is introduced. With the help of this package it is possible to keep microdata confidential in a very effective way. The concept is thoroughly explained and its application is demonstrated using real-world data. In addition to that, the robustification of disclosure methods is described. Many SDC-methods for microdata developed so far can be influenced by outliers to a great extent resulting in a high loss of information of the perturbed data. Missing values are the second topic of this book. The application of visualisation tools for the analysis of missing values, preceding the choice of an imputation method, is highlighted. In addition to that, new methods for the imputation of composition data are introduced. Due to the linear dependence of the variables from compositional data, reasonalbe imputations can be made by considering the special nature of such data.