BizRel is a business relation dataset composed of 10k instances in english and another 10k instances in french. It has been compiled to train relation extraction models at the sentence level. Each relation instance is a tuple (Sentence, entity1, entity2, relation_type), where:
- entity1 and entity2 are named entities of type Organization (ORG).
- sentences are raw sentences collected from the web.
- relation_types are relation labels manually annotated by crowdsourcing (using the Isahit platform).
A part of this dataset can be used for research purposes. Commercial use is not allowed as well as third party distribution.
@inproceedings{khaldi-etal-2022-hows,
title = "How{'}s Business Going Worldwide ? A Multilingual Annotated Corpus for Business Relation Extraction",
author = "Khaldi, Hadjer and
Benamara, Farah and
Pradel, Camille and
Sigel, Gr{\'e}goire and
Aussenac-Gilles, Nathalie",
booktitle = "Proceedings of the Thirteenth Language Resources and Evaluation Conference",
month = jun,
year = "2022",
address = "Marseille, France",
publisher = "European Language Resources Association",
url = "https://aclanthology.org/2022.lrec-1.394",
pages = "3696--3705",}
To have access to the dataset, please fill this form. For further information, please contact hadjer@geotrend.fr