Microaggregation for Categorical Variables: A Median Based Approach

摘要

Microaggregation is a masking procedure used for protecting confidential data prior to their public release. This technique, that relies on clustering and aggregation techniques, is solely used for numerical data. In this work we introduce a microaggregation procedure for categorical variables. We describe the new masking method and we analyse the results it obtains according to some indices found in the literature. The method is compared with Top and Bottom Coding, Global receding, Rank Swapping and PRAM.