Entity resolution on-demand

Entity Resolution (ER) aims to identify and merge records that refer the same real-world entity. ER is typically employed as an expensive cleaning step on entire data before consuming it. Yet, determining which entities are useful once cleaned depends solely user's application, may need only a fraction of them. For instance, when dealing with Web data, we would like be able filter interest gathered from multiple sources without entire, continuously-growing data. Similarly, querying lakes, want transform on-demand return results in timely manner---a fundamental requirement ELT ( Extract-Load-Transform ) pipelines. We propose BrewER , framework evaluate SQL SP queries dirty while progressively returning if they were issued tries focus effort one entity at time, following ORDER BY predicate. Thus, it inherently supports top-k stop-and-resume execution. wide range applications, significant amount resources can saved. exhaustively show efficacy four datasets.