2010/08/26 by Jaffer Gardezi, Gardezi, Jaffer, Leopoldo Bertossi +3
Decision Sciences · Computer Science · #Data Quality and Management #Advanced Database Systems and Queries #Semantic Web and Ontologies
paper · pdf · doi:10.48550/arxiv.1008.4627
Matching dependencies (MDs) were introduced to specify the identification or\nmatching of certain attribute values in pairs of database tuples when some\nsimilarity conditions are satisfied. Their enforcement can be seen as a natural\ngeneralization of entity resolution. In what we call the "pure case" of MDs,\nany value from the underlying data domain can be used for the value in common\nthat does the matching. We investigate the semantics and properties of data\ncleaning through the enforcement of matching dependencies for the pure case. We\ncharacterize the intended clean instances and also the "clean answers" to\nqueries as those that are invariant under the cleaning process. The complexity\nof computing clean instances and clean answers to queries is investigated.\nTractable and intractable cases depending on the MDs and queries are\nidentified. Finally, we establish connections with database "repairs" under\nintegrity constraints.\n