Query-Oriented Data Cleaning with Oracles

摘要: As key decisions are often made based on information contained in a database, it is important for the database to be as complete and correct possible. For this reason, many data cleaning tools have been developed automatically resolve inconsistencies databases. However, provide only best-effort results usually cannot eradicate all errors that may exist database. Even more importantly, existing do not typically address problem of determining what missing from To overcome limitations techniques, we present QOCO, novel query-oriented system with oracles. Under framework, incorrect (resp. missing) tuples removed (added to) result query through edits applied underlying where derived by interacting domain experts which model oracle crowds. We show minimal interactions crowds derive removing (adding) (missing) NP-hard general heuristic algorithms interact Finally, implement our prototype QOCO effective efficient comprehensive suite experiments.