Uncertainty in data integration: current approaches and open problems

Uncertainty is an intrinsic feature of automatic and semiautomatic data integration processes. Although many solutions have been proposed to reduce uncertainty, if we do not explicitly represent and keep it up to the end of the integration process we risk to lose relevant information, and to produce misleading results. Models for uncertain data can then be used to represent integrated data sources resulting from uncertain data integration processes. In this paper we present a survey of existing approaches directly dealing with uncertainty in data integration, define a generic data integration process that explicitly represents uncertainty during all its steps, and present some preliminary results and open issues in the field.