The viewer is disabled because this dataset repo requires arbitrary Python code execution. Please consider removing the loading script and relying on automated data support (you can use convert_to_parquet from the datasets library). If this is not possible, please open a discussion for direct help.

Dataset Card for MasakhaNER

Dataset Summary

MasakhaNER is the first large publicly available high-quality dataset for named entity recognition (NER) in ten African languages.

Named entities are phrases that contain the names of persons, organizations, locations, times and quantities. Example:

[PER Wolff] , currently a journalist in [LOC Argentina] , played with [PER Del Bosque] in the final years of the seventies in [ORG Real Madrid] .

MasakhaNER is a named entity dataset consisting of PER, ORG, LOC, and DATE entities annotated by Masakhane for ten African languages:

  • Amharic
  • Hausa
  • Igbo
  • Kinyarwanda
  • Luganda
  • Luo
  • Nigerian-Pidgin
  • Swahili
  • Wolof
  • Yoruba

The train/validation/test sets are available for all the ten languages.

For more details see https://arxiv.org/abs/2103.11811

Supported Tasks and Leaderboards

[More Information Needed]

  • named-entity-recognition: The performance in this task is measured with F1 (higher is better). A named entity is correct only if it is an exact match of the corresponding entity in the data.

Languages

There are ten languages available :

  • Amharic (amh)
  • Hausa (hau)
  • Igbo (ibo)
  • Kinyarwanda (kin)
  • Luganda (kin)
  • Luo (luo)
  • Nigerian-Pidgin (pcm)
  • Swahili (swa)
  • Wolof (wol)
  • Yoruba (yor)

Dataset Structure

Data Instances

The examples look like this for Yorùbá:

from datasets import load_dataset
data = load_dataset('masakhaner', 'yor') 

# Please, specify the language code

# A data point consists of sentences seperated by empty line and tab-seperated tokens and tags. 
{'id': '0',
 'ner_tags': [B-DATE, I-DATE, 0, 0, 0, 0, 0, B-PER, I-PER, I-PER, O, O, O, O],
 'tokens': ['Wákàtí', 'méje', 'ti', 'ré', 'kọjá', 'lọ', 'tí', 'Luis', 'Carlos', 'Díaz', 'ti', 'di', 'awati', '.']
}

Data Fields

  • id: id of the sample
  • tokens: the tokens of the example text
  • ner_tags: the NER tags of each token

The NER tags correspond to this list:

"O", "B-PER", "I-PER", "B-ORG", "I-ORG", "B-LOC", "I-LOC", "B-DATE", "I-DATE",

In the NER tags, a B denotes the first item of a phrase and an I any non-initial word. There are four types of phrases: person names (PER), organizations (ORG), locations (LOC) and dates & time (DATE).

It is assumed that named entities are non-recursive and non-overlapping. In case a named entity is embedded in another named entity usually, only the top level entity is marked.

Data Splits

For all languages, there are three splits.

The original splits were named train, dev and test and they correspond to the train, validation and test splits.

The splits have the following sizes :

Language train validation test
Amharic 1750 250 500
Hausa 1903 272 545
Igbo 2233 319 638
Kinyarwanda 2110 301 604
Luganda 2003 200 401
Luo 644 92 185
Nigerian-Pidgin 2100 300 600
Swahili 2104 300 602
Wolof 1871 267 536
Yoruba 2124 303 608

Dataset Creation

Curation Rationale

The dataset was introduced to introduce new resources to ten languages that were under-served for natural language processing.

[More Information Needed]

Source Data

The source of the data is from the news domain, details can be found here https://arxiv.org/abs/2103.11811

Initial Data Collection and Normalization

The articles were word-tokenized, information on the exact pre-processing pipeline is unavailable.

Who are the source language producers?

The source language was produced by journalists and writers employed by the news agency and newspaper mentioned above.

Annotations

Annotation process

Details can be found here https://arxiv.org/abs/2103.11811

Who are the annotators?

Annotators were recruited from Masakhane

Personal and Sensitive Information

The data is sourced from newspaper source and only contains mentions of public figures or individuals

Considerations for Using the Data

Social Impact of Dataset

[More Information Needed]

Discussion of Biases

[More Information Needed]

Other Known Limitations

Users should keep in mind that the dataset only contains news text, which might limit the applicability of the developed systems to other domains.

Additional Information

Dataset Curators

Licensing Information

The licensing status of the data is CC 4.0 Non-Commercial

Citation Information

Provide the BibTex-formatted reference for the dataset. For example:

@article{Adelani2021MasakhaNERNE,
  title={MasakhaNER: Named Entity Recognition for African Languages},
  author={D. Adelani and Jade Abbott and Graham Neubig and Daniel D'Souza and Julia Kreutzer and Constantine Lignos 
  and Chester Palen-Michel and Happy Buzaaba and Shruti Rijhwani and Sebastian Ruder and Stephen Mayhew and 
  Israel Abebe Azime and S. Muhammad and Chris C. Emezue and Joyce Nakatumba-Nabende and Perez Ogayo and 
  Anuoluwapo Aremu and Catherine Gitau and Derguene Mbaye and J. Alabi and Seid Muhie Yimam and Tajuddeen R. Gwadabe and
  Ignatius Ezeani and Rubungo Andre Niyongabo and Jonathan Mukiibi and V. Otiende and Iroro Orife and Davis David and 
  Samba Ngom and Tosin P. Adewumi and Paul Rayson and Mofetoluwa Adeyemi and Gerald Muriuki and Emmanuel Anebi and 
  C. Chukwuneke and N. Odu and Eric Peter Wairagala and S. Oyerinde and Clemencia Siro and Tobius Saul Bateesa and 
  Temilola Oloyede and Yvonne Wambui and Victor Akinode and Deborah Nabagereka and Maurice Katusiime and 
  Ayodele Awokoya and Mouhamadane Mboup and D. Gebreyohannes and Henok Tilaye and Kelechi Nwaike and Degaga Wolde and
   Abdoulaye Faye and Blessing Sibanda and Orevaoghene Ahia and Bonaventure F. P. Dossou and Kelechi Ogueji and 
   Thierno Ibrahima Diop and A. Diallo and Adewale Akinfaderin and T. Marengereke and Salomey Osei},
  journal={ArXiv},
  year={2021},
  volume={abs/2103.11811}
}

Contributions

Thanks to @dadelani for adding this dataset.

Downloads last month
1,532

Models trained or fine-tuned on masakhane/masakhaner