Ontology-based data quality framework for data stream applications

Data Stream Management Systems (DSMS) have been proposed to address the challenges of applications which produce continuous, rapid streams of data that have to be processed in real-time. Data quality (DQ) plays an important role in DSMS as there is usually a trade-off between accuracy and consistency on the one hand, and timeliness and completeness on the other hand. Previous work on data quality in DSMS has focused only on specific aspects of DQ. In this paper, we present a flexible, holistic ontology-based data quality framework for data stream applications. Our DQ model is based on a threefold notion of DQ. First, content-based evaluation of DQ uses semantic rules which can be userdefined in an extensible ontology. Second, query-based evaluation adds DQ information to the query results and updates it while queries are being processed. Third, the application-based evaluation can use any kind of function which computes an application-specific DQ value. The whole DQ process is driven by the metadata managed in an ontology which provides a semantically clear definition of the DQ features of the DSMS. The evaluation of our approach in two case studies in the domain of traffic information systems has shown that our framework provides the required flexibility, extensibility, and performance for DQ management in DSMS.