A Model for Improving Guilt Probabilities to Identify Data Leackages

Some Times Sensitive Data may be distributes by a data distributor to a set of trusted agents (third parties), But Some of the distributed data we may find in an unauthorized website or systems. So the distributor must assess the probability of data leakage from its trusted agents. So in this paper we propose a model for assessing the guilt probability of agents, in a way that improves the chances of identifying a data leakage. We also propose algorithms for distributing objects to agents. Finally we also consider the option of adding fake data objects to the distributed data set. Such objects do not correspond to the real entities but appear realistic to the trusted agents. These fake objects used like watermarking for the entire dataset. These schemes do not make any alternations of the released data