grobid/dataseer repository overview

dataseer-ml is a Machine Learning tool aiming at identifying implicit mentions of datasets in a scientific article and to classify the identified datasets in a hierarchy of dataset types, these data types being directly derived from MeSH. Most of the datasets discussed in scientific articles are actually not named, but these data are part of the disclosed scientific work and should be shared properly to meet the FAIR requirements.

The module can process a variety of scientific article formats, including mainstream publisher's native XML submission formats: PDF, TEI, JATS/NLM, ScholarOne, BMJ, Elsevier staging format, OUP, PNAS, RSC, Sage, Wiley, etc.