README.md

metadata

license: cc-by-sa-4.0
task_categories:
  - token-classification
language:
  - sr
pretty_name: SrpELTeC-gold-NER training dataset
size_categories:
  - 100K<n<1M
tags:
  - NER

Named Entity Recognition Training corpus for Serbian – The selection of 11 full novels and excerpts from 15 novels from Serbian literary corpus of novels written more than a century ago, have been automatically labelled with SrpNER system for Serbian in the first stage of the gold standard preparation. Contains 330.119 tokens, 7 classes: person, organization, location, event, work, demonym, role.

It is constituted of a single jsonl file that can be loaded via:

from datasets import load_dataset
dataset = load_dataset("jerteh/SrpELTeC-gold-NER")

Preview:

ds = dataset["train"][36042]
for x, y, z in zip(ds["token"], ds["lemma"], ds["ner"]):
    print(x, y, z)


Setila setiti O
se se O
da da O
je jesam O
Petrovski Petrovski B-EVENT
post post I-EVENT
. . O

Citation:

@inproceedings{frontini2020named,
  title={Named entity recognition for distant reading in ELTeC},
  author={Frontini, Francesca and Brando, Carmen and Byszuk, Joanna and Galleron, Ioana and Santos, Diana and Stankovi{\'c}, Ranka},
  booktitle={CLARIN Annual Conference 2020},
  year={2020}
}