Permalink
Switch branches/tags
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
5 lines (2 sloc) 1.26 KB

ZBW's RAS-GND author ID mapping

ZBW's EconBiz search portal contains metadata about publications in economics from multiple sources. Therefore, duplicates occur in the data (e.g. a publication as entered in RePEc and as entered in ZBW's catalogue). In order to detect such duplicates, a project has been launched. The goal of the project is to not only find duplicate publications but also resolve different author ids that refer to the same person, since in the ZBW catalogue the authors of publications are identified by their GND ID, and in RePEc by their RAS ID. Hence, in a first iteration we derived a mapping of authors from the matching papers. That has been done very conservatively, with a threshold of more than 10 matching papers. However, due to potentially frequent co-authorship of authors, we still need to constrain the matched authors to have similar names. Due to the previous constraint we can apply a rather weak constraint (normalized Levenshtein distance < .5) on the author names to further increase the precision of the matching with only a relatively small decrease in recall. That yielded 3,081 GND/RAS mappings.