Uma abordagem de privacidade diferencial para consultas sobre dados RDF no contexto de redes sociais.

Abstract

As the amount of collected information in RDF format grows, the development of solutions for privacy of individuals, their attributes and relationships with others become a more important subject of study. However, privacy solutions are not well suitable for this specific type of data, because they usually do not consider relationships between individuals, which are crucial to semantic data and social networks. Although differential privacy is the most suitable technique for statistical queries, there is still work to be done in this context. This paper presents two main contributions for privacy preserving statistic queries with a relationship as a filter. The first one describes a complete approach to apply ε-differential privacy for linked data and the second one presents an auxiliary data structure and algorithms to efficiently compute parameters for the differential privacy mechanism, i.e. the query’s actual value and sensitivity of the data for the given query. We conclude by evaluating our contributions, in real data, presenting utility analysis considering different values of ε as well as performance analysis of our data structure and algorithms.