KATARA

TLDR

Classical data‑cleaning methods rely on integrity constraints, statistics, or machine learning, but their accuracy is limited and can be improved by consulting master data and experts, while knowledge bases and crowdsourcing now offer scalable ways to achieve higher accuracy. KATARA is introduced as a knowledge‑base and crowd‑powered system that interprets table semantics, aligns data with a KB, detects errors, and produces top‑k repair suggestions. It operates by mapping table semantics to a knowledge base, flagging inconsistencies, and leveraging crowd input to generate ranked repair candidates. Experiments demonstrate that KATARA can be applied to diverse datasets and knowledge bases, efficiently annotating data and proposing repair suggestions.

Abstract

Classical approaches to clean data have relied on using integrity constraints, statistics, or machine learning. These approaches are known to be limited in the cleaning accuracy, which can usually be improved by consulting master data and involving experts to resolve ambiguity. The advent of knowledge bases KBs both general-purpose and within enterprises, and crowdsourcing marketplaces are providing yet more opportunities to achieve higher accuracy at a larger scale. We propose KATARA, a knowledge base and crowd powered data cleaning system that, given a table, a KB, and a crowd, interprets table semantics to align it with the KB, identifies correct and incorrect data, and generates top-k possible repairs for incorrect data. Experiments show that KATARA can be applied to various datasets and KBs, and can efficiently annotate data and suggest possible repairs.