Named entity recognition (NER) on historical DBNL data

The dataset contains the public domain part of DBNL (3542publications) annotated with 3 NER models. The models were fine-tuned on Dutch historical data and described in this paper:https://aclanthology.org/2024.lt4hala-1.4/ . The dataset consists of 4parts: the predictions of each of the 3 NER models, and thepredictions obtained by combining the models in majority voting. It canbe used as silver data for training entity linking models, or in digitalhumanities applications. The annotations are available as tsv files inthe standard IOB format (each token is labelled as beginning, insideor outside of a named entity).Acknowledgements: support with the data and research was providedby Marieke Moolenaar.