The Theory of Probabilistic Databases 论文

摘要

ABS!l’RACT A theory of probabilistic databases is outlined. This theory is one component of an integrated approach to data-modelling that accomodates both probabilistic and relational data. In fact, many of the results presented here were developed in the context of a framework for structural modelling of systems. Much that is fundamental to relational database theory was also developed in this context, and previous to the introduction by Codd of the relational model of data. Probabilistic databases can store types of information that cannot be represented using the relational model. Probabilistic databases may also be viewed as generalisations of relational databases; any relational database can be represented without loss of information by a probabilistic database. A number of relational database concepts are shown to have probabilistic counterparts. In many cases, it is preferable to deal with the probabilistic formulation of a concept even when applying it to a relational database. For example, we define a new project-join mapping for relational databases that is based on transf orming a relational to a probabilistic database. This mapping is shown to have more fmed points than the standard one.