Deduplication on knihovny.cz portal

Summary: The aim of this article is to outline the processing of metadata on Knihovny.cz portal, and in particular to present the process of deduplication, i.e. the recognition and clustering of records related to the same document Knihovny.cz portal deduplicates by creating merged records to link local records from individual metadata sources. Deduplication consists of a sequence of steps comparing different combinations of deduplication keys. The paper further describes the ways in which the portal aims to minimize complications caused by incorrect metadata, the weaknesses and strengths of the chosen methods and the results of their application on records from all involved sources. It concludes with a discussion of assumptions for improving the current state and possibilities for further development.

Keywords: bibliographic records de­‑duplication, MARC 21, Knihovny.cz portal

Mgr. Jana Kurfürstová, Mgr. Petra Žabičková, Bc. Tomáš Cejpek / Moravská zemská knihovna v Brně (Moravian Library), Kounicova 65a, 601 87 Brno