A Bayesian approach for on-line max auditing of dynamic statistical databases

摘要

In this paper we propose a method for on-line max auditing of dynamic statistical databases. The method extends the Bayesian approach presented in [2], [3] and [4] for static databases. A Bayesian network addresses disclosures based on probabilistic inferences that can be drawn from released data; we have developed algorithms to update the network whenever the database changes. In particular, we consider the case in which records are added or deleted, or some sensitive values change their value. The paper introduces the algorithms and discusses results of a preliminary set of of experimental trials.