TeCla (Text Classification) Catalan dataset

BibTeX citation

If you use any of these resources (datasets or models) in your work, please cite our latest paper:

@inproceedings{armengol-estape-etal-2021-multilingual,
    title = "Are Multilingual Models the Best Choice for Moderately Under-resourced Languages? {A} Comprehensive Assessment for {C}atalan",
    author = "Armengol-Estap{\'e}, Jordi  and
      Carrino, Casimiro Pio  and
      Rodriguez-Penagos, Carlos  and
      de Gibert Bonet, Ona  and
      Armentano-Oller, Carme  and
      Gonzalez-Agirre, Aitor  and
      Melero, Maite  and
      Villegas, Marta",
    booktitle = "Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021",
    month = aug,
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.findings-acl.437",
    doi = "10.18653/v1/2021.findings-acl.437",
    pages = "4933--4946",
}

Digital Object Identifier (DOI) and access to dataset files

https://doi.org/10.5281/zenodo.4627198

Introduction

TeCla is a Catalan News corpus for thematic Text Classification tasks. It contains 153.265 articles classified under 30 different categories.

The source data is crawled from the ACN (Catalan News Agency) site: [ http://www.acn.cat] , and used under CC-BY-NC-ND 4.0 licence. The dataset is released under the same licence, and is intended exclusively for training Machine Learning models.

This dataset was developed by BSC TeMU as part of the AINA project, and intended as part of CLUB (Catalan Language Understanding Benchmark). It is part of the Catalan Language Understanding Benchmark (CLUB) as presented in:

Armengol-Estapé J., Carrino CP., Rodriguez-Penagos C., de Gibert Bonet O., Armentano-Oller C., Gonzalez-Agirre A., Melero M. and Villegas M.,Are Multilingual Models the Best Choice for Moderately Under-resourced Languages? A Comprehensive Assessment for Catalan". Findings of ACL 2021 (ACL-IJCNLP 2021).

Supported Tasks and Leaderboards

Text classification, Language Model

Languages

CA- Catalan

Directory structure

• .gitattributes
• README.md
• dev.json - json-formatted file with the dev split of the dataset
• tecla.py
• test.json - json-formatted file with the test split of the dataset
• train.json - json-formatted file with the train split of the dataset

Dataset Structure

Data Instances

Three json files, one for each split.

Data Fields

We used a simple model with the article text and associated labels, without further metadata.

Example:

{"version": "1.0",
 "data":
   [
    {
     'sentence': 'L\\\\'editorial valenciana Media Vaca, Premi Nacional a la Millor Tasca Editorial Cultural del 2018. El jurat en destaca la cura "exquisida" del catàleg, la qualitat dels llibres i el "respecte" pels lectors. ACN Madrid.-L\\\\'editorial valenciana Media Vaca ha obtingut el Premi Nacional a la Millor Labor Editorial Cultural corresponent a l\\\\'any 2018 que atorga el Ministeri de Cultura i Esports. El guardó pretén distingir la tasca editorial d\\\\'una persona física o jurídica que hagi destacat per l\\\\'aportació a la vida cultural espanyola. El premi és de caràcter honorífic i no té dotació econòmica. En el cas de Media Vaca, fundada pel valencià Vicente Ferrer i la bilbaïna Begoña Lobo, el jurat n\\\\'ha destacat la cura "exquisida" del catàleg, la qualitat dels llibres i el "respecte" pels lectors i per la resta d\\\\'agents de la cadena del llibre. Media Vaca va publicar els primers llibres el desembre del 1998. El catàleg actual el componen 64 títols dividits en sis col·leccions, que barregen ficció i no ficció. Des del Ministeri de Cultura es destaca que la il·lustració té un pes "fonamental" als productes de l\\\\'editorial i que la majoria de projectes solen partir de propostes literàries i textos preexistents. L\\\\'editorial ha rebut quatre vegades el Bologna Ragazzi Award. És l\\\\'única editorial estatal que ha aconseguit el guardó que atorga la Fira del Llibre per a Nens de Bolonya, la més important del sector.', 
    'label': 'Lletres'
    },
    .
    .
    .
  ]
}

Data Splits

• train.json: 122587 article-label pairs
• dev.json: 15339 article-label pairs
• test.json: 15339 article-label pairs

Labels

'Societat', 'Política', 'Turisme', 'Salut', 'Economia', 'Successos', 'Partits', 'Educació', 'Policial', 'Medi ambient', 'Parlament', 'Empresa', 'Judicial', 'Unió Europea', 'Comerç', 'Cultura', 'Cinema', 'Govern', 'Lletres', 'Infraestructures', 'Música', 'Festa i cultura popular', 'Teatre', 'Mobilitat', 'Govern espanyol', 'Equipaments i patrimoni', 'Meteorologia', 'Treball', 'Trànsit', 'Món'

Labels in the dataset by frequency

train.json: 122587 articles

Label	Num art	% art
Societat	24975	20.37%
Política	18344	14.96%
Partits	10056	8.2%
Successos	7874	6.42%
Judicial	5788	4.72%
Policial	5557	4.53%
Salut	5430	4.43%
Economia	5032	4.1%
Parlament	4176	3.41%
Medi_ambient	3027	2.47%
Música	2872	2.34%
Educació	2757	2.25%
Empresa	2698	2.2%
Cultura	2495	2.04%
Unió_Europea	2064	1.68%
Govern	2039	1.66%
Infraestructures	1740	1.42%
Treball	1655	1.35%
Mobilitat	1624	1.32%
Cinema	1560	1.27%
Teatre	1492	1.22%
Turisme	1232	1.01%
Equipaments_i_patrimoni	1229	1.0%
Lletres	1180	0.96%
Meteorologia	1080	0.88%
Comerç	984	0.8%
Govern_espanyol	983	0.8%
Món	893	0.73%
Festa_i_cultura_popular	888	0.72%
Trànsit	863	0.7%

dev.json and test.json: 153265 articles each split

Label	Num art	% art
Societat	3122	20.35%
Política	2294	14.96%
Partits	1257	8.19%
Successos	985	6.42%
Judicial	724	4.72%
Policial	695	4.53%
Salut	679	4.43%
Economia	630	4.11%
Parlament	523	3.41%
Medi_ambient	379	2.47%
Música	359	2.34%
Educació	345	2.25%
Empresa	338	2.2%
Cultura	312	2.03%
Unió_Europea	258	1.68%
Govern	256	1.67%
Infraestructures	218	1.42%
Treball	208	1.36%
Mobilitat	204	1.33%
Cinema	195	1.27%
Teatre	187	1.22%
Turisme	154	1.0%
Equipaments_i_patrimoni	154	1.0%
Lletres	148	0.96%
Meteorologia	135	0.88%
Govern_espanyol	124	0.81%
Comerç	123	0.8%
Festa_i_cultura_popular	112	0.73%
Món	112	0.73%
Trànsit	109	0.71%

Dataset Creation

Methodology

We crawled 219.586 articles from the Catalan News Agency ( www.acn.cat ) newswire archive, the latest from October 11, 2020. We used the "subsection" category as a classification label, after excluding territorial labels (see territorial_labels.txt file) and labels with less than 2000 occurrences. With this criteria compiled a total of 153.265 articles for this text classification dataset.

Curation Rationale

We used the "subsection" category as a classification label, after excluding territorial labels (see territorial_labels.txt file) and labels with less than 2000 occurrences.

Source Data

Initial Data Collection and Normalization

The source data are crawled articles from ACN (Catalan News Agency) site: www.acn.cat

Who are the source language producers?

The Catalan News Agency (CNA, in Catalan: Agència Catalana de Notícies (ACN)) is a news agency owned by the Catalan government via the public corporation Intracatalònia, SA. It is one of the first digital news agencies created in Europe and has been operating since 1999 (source: [ https://en.wikipedia.org/wiki/Catalan_News_Agency] )

Annotations

Annotation process

We used the "subsection" category as a classification label, after excluding territorial labels (see territorial_labels.txt file) and labels with less than 2000 occurrences.

Who are the annotators?

Editorial staff classified the articles under the different thematic sections, and we extracted these from metadata.

Dataset Curators

Casimiro Pio Carrino, Carlos Rodríguez and Carme Armentano, from BSC-CNS

Personal and Sensitive Information

No personal or sensitive information included.

Considerations for Using the Data

Social Impact of Dataset

[More Information Needed]

Discussion of Biases

[More Information Needed]

Other Known Limitations

[More Information Needed]

Contact

Carlos Rodríguez-Penagos or Carme Armentano-Oller ( bsc-temu@bsc.es )

License

This work is licensed under a Attribution-NonCommercial-NoDerivatives 4.0 International License .