metadata

annotations_creators:
  - crowdsourced
language_creators:
  - found
language:
  - he
license:
  - other
multilinguality:
  - monolingual
size_categories:
  - 10K<n<100K
source_datasets:
  - extended|other-reuters-corpus
task_categories:
  - token-classification
task_ids:
  - named-entity-recognition
train-eval-index:
  - config: bmc
    task: token-classification
    task_id: entity_extraction
    splits:
      train_split: train
      eval_split: validation
      test_split: test
    col_mapping:
      tokens: tokens
      ner_tags: tags
    metrics:
      - type: seqeval
        name: seqeval

Splits for the Ben-Mordecai and Elhadad Hebrew NER Corpus (BMC)

In order to evaluate performance in accordance with the original Ben-Mordecai and Elhadad (2005) work, we provide three 75%-25% random splits.

Only the 7 entity categories viable for evaluation were kept (DATE, LOC, MONEY, ORG, PER, PERCENT, TIME) --- all MISC entities were filtered out.
Sequence label scheme was changed from IOB to BIOES
The dev sets are 10% taken out of the 75%

Citation

If you use use the BMC corpus, please cite the original paper as well as our paper which describes the splits:

Ben-Mordecai and Elhadad (2005):

@mastersthesis{naama,
  title={Hebrew Named Entity Recognition},
  author={Ben-Mordecai, Naama},
  advisor={Elhadad, Michael},
  year={2005},
  url="https://www.cs.bgu.ac.il/~elhadad/nlpproj/naama/",
  institution={Department of Computer Science, Ben-Gurion University},
  school={Department of Computer Science, Ben-Gurion University},
}

Bareket and Tsarfaty (2020)

@misc{bareket2020neural,
      title={Neural Modeling for Named Entities and Morphology (NEMO^2)}, 
      author={Dan Bareket and Reut Tsarfaty},
      year={2020},
      eprint={2007.15620},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}