BUSiness Transaction Entity Recognition dataset.
BUSTER is an Entity Recognition (ER) benchmark for entities related to business transactions. It consists of a gold corpus of 3779 manually annotated documents on financial transactions that were randomly divided into 5 folds, plus an additional silver corpus of 6196 automatically annotated documents that were created by the model-optimized RoBERTa system.
Paper | Code | Results | Date | Stars |
---|