Tuple Merging in Probabilistic Databases

Real-world data are often uncertain and incomplete. In probabilistic relational data models uncertainty can be modeled on two levels. First by representing the uncertain instance of a tuple by a set of possible instances and second by assigning each tuple with its degree of membership to the considered relation. To overcome incompleteness, data from multiple sources need to be combined. In order to combine data from autonomous probabilistic databases, an integration of probabilistic data has to be performed. Until now, however, data integration approaches have focused on the integration of certain source data (relational or XML). There has been only less attention on the integration of uncertain (esp. probabilistic) source data so far. In this paper, we consider probabilistic tuple merging being an essential step in the integration of probabilistic data. We present techniques for merging uncertain instance data as well as for merging different degrees of tuple membership.