Connecting authors to their work – the challenges and successes

Abstract

A key element of any CRIS is the tracking of research output in the form of publications - and associating them to the researchers who authored them. Creating a comprehensive list – retrospectively and for ongoing new research - is not a simple task and often institutions rely on manual work. There are multiple challenges from lack of or inconsistent metadata to missing identifiers. In this session we are discussing the use of algorithms, developed with machine learning methodologies, for automating the author-work associating at scale. Using the example of Smart Harvesting in Esploro, we outline the challenges and the successes of such an approach.