ANON-a flexible tool for achieving opti...

Anonymization of data is an indispensable task whensensible data has to be shared between different organizations, e.g. when sharinggenomic data gathered at and shared between different medical researchorganizations. Several concepts have been discussed during the last coupleyears, for instance k-anonymity and l-diversity. Both conceptsare privacy constraints that must be guaranteed in query results to assure ahigh privacy degree. However, the approaches presented so far cannot deal withNULL values in the underlying data sources. In this technical report we willdiscuss the problem of NULL values for anonymization in detail and present anddiscuss our approach for k-anonymity which explicitly considers NULLvalues. Furthermore, we will present our implementation of this approach.