GGPONC 2.0 (Coordination Ellipses)

Dataset of more than 4,000 manually annotated elliptical coordinated compound noun phrases (ECCNPs) in German medical text (subset of GGPONC 2.0), along with the respective ground truth resolutions.

Published in: Resolving Elliptical Compounds in German Medical Text @ BioNLP'23 (ACL Workshop)

Files:

ggponc_ellipses_compounds.tsv (examples of ECCNPs + resolution)

ggponc_no_ellipses_small.tsv (manually validated subset of sentences not containing ECCNPs)

ggponc_no_ellipses_all.tsv (more extensive, but potentially noisy subset of sentences not containing ECCNPs)

ellipses_splits.csv (assignment of files to train / dev / test splits)