Normalização de nomes de autores em fontes de informação institucionais: proposta de um método automático de verificação de erros

TL;DRAbstract

The recovery of scientific literature by authoring is a challenge for many maintainers of databases, due to the ambiguity caused by problems originated from lack of control at the time of indexing. This paper presents an automatic approach of checking for errors in authorship metadata of University of São Paulo scientific production database (Dedalus) comparing these data with data recovered from the human resources database. Using approximate string matching algorithms, these data from human resources is compared with the scientific production data of four institutes from USP (covering the period 2006-2010). Based on this pilot study it was possible to establish interoperability mechanisms between Dedalus database and the USP human resources database. As an immediate result, it was possible to map the percentage of errors and to create mechanisms of interference, establish a timeline to expand the study to other institutes from USP, and standardization procedures.

Chat with Paper

AI Agents for this Paper