# KazNERD: Kazakh Named Entity Recognition Dataset

This repository contains the Kazakh Named Entity Recognition Dataset (KazNERD), annotation guidelines, and code for training various Named Entity Recognition (NER) models.  KazNERD addresses the scarcity of publicly available annotated Kazakh corpora.

## Dataset Summary

KazNERD is a corpus of 112,702 sentences extracted from television news text, annotated with 136,333 named entities across 25 entity classes.  Annotation was performed by two native Kazakh speakers, supervised by a researcher, using the IOB2 scheme. The data is provided in the CoNLL 2002 format.

##  Statistics

| Statistic          | Value       |
|----------------------|-------------|
| Number of Sentences | 112,702     |
| Number of Annotations | 136,333     |
| Number of Entity Classes | 25          |
| Annotation Scheme   | IOB2        |
| Language            | Kazakh      |


##  Models

The repository includes code for training NER models using CRF, BiLSTM-CNN-CRF, BERT, and XLM-RoBERTa. The best-performing model achieved an exact match F1-score of 97.22% on the test set.

## Citation

```bibtex
@inproceedings{yeshpanov-etal-2022-kaznerd,
    title = "{K}az{NERD}: {K}azakh Named Entity Recognition Dataset",
    author = "Yeshpanov, Rustem  and
      Khassanov, Yerbolat  and
      Varol, Huseyin Atakan",
    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-1.44",
    pages = "417--426",
    abstract = "We present the development of a dataset for Kazakh named entity recognition. The dataset was built as there is a clear need for publicly available annotated corpora in Kazakh, as well as annotation guidelines containing straightforward{---}but rigorous{---}rules and examples. The dataset annotation, based on the IOB2 scheme, was carried out on television news text by two native Kazakh speakers under the supervision of the first author. The resulting dataset contains 112,702 sentences and 136,333 annotations for 25 entity classes. State-of-the-art machine learning models to automatise Kazakh named entity recognition were also built, with the best-performing model achieving an exact match F1-score of 97.22{\%} on the test set. The annotated dataset, guidelines, and codes used to train the models are freely available for download under the CC BY 4.0 licence from https://github.com/IS2AI/KazNERD.",
}
```

[https://github.com/IS2AI/KazNERD](https://github.com/IS2AI/KazNERD)