Tractability in probabilistic databases

Abstract

Probabilistic databases are motivated by a large and diverse set of applications that need to query and process uncertain data. Uncertain and probabilistic data arises in RFID systems [22], information extraction [12], data cleaning [1], scientific data management [17], biomedical data integration [9], business intelligence [14], approximate schema mappings [10], data deduplication [13]. All these applications have large collections of data, where some, or most individual data items are uncertain.