An FRBR-based approach for transforming MARC records into linked data

In this study, a practical workflow is outlined for transforming MARC records into Linked Data according to classes and properties of FRBR and FRAD defined in the RDA Registry. Four bibliographic records of ‘Pride and Prejudice’ were selected as subject. One was in English and the others were Chinese records to illustrate the hidden relationships embedded in MARC records, including translation, version and reproduction, relationships between classes in FRBR group 1, and relationships between FRBR group 1 and FRAD for Linked Data. In total, the proposed workflow is composed of three stages and five tasks as follows: data preparation and selection (comprising two tasks, i.e., changing MARC tags into semantic labels with their data and selecting significant labels and their data for Linked Data), data modeling (comprising two tasks, i.e., aligning MARC labels and their data with semantic equivalent classes defined in the RDA Registry and building the relationships between selected classes through the properties of the RDA Registry), and reusing existing value vocabularies for external linking (comprising one task, such as linking terms from VIAF, TGN and so on). Issues related to de-duplication and collaboration of LD are also addressed.