projecte-aina/ancora-ca-ner
收藏资源简介:
--- annotations_creators: - expert-generated language_creators: - found language: - ca license: - cc-by-4.0 multilinguality: - monolingual pretty_name: ancora-ca-ner size_categories: - unknown source_datasets: [] task_categories: [] task_ids: [] --- # Dataset Card for AnCora-Ca-NER ## Dataset Description - **Website:** https://zenodo.org/records/4761746 - **Homepage** [Projecte AINA](https://projecteaina.cat/tech/) - **Repository** [HuggingFace](https://huggingface.co/projecte-aina) - **Paper:** [Are Multilingual Models the Best Choice for Moderately Under-resourced Languages? A Comprehensive Assessment for Catalan](https://arxiv.org/abs/2107.07903) - **Paper:** [AnCora: Multilevel Annotated Corpora for Catalan and Spanish](http://www.lrec-conf.org/proceedings/lrec2008/pdf/35_paper.pdf) - **Point of Contact** langtech@bsc.es ### Dataset Summary This is a dataset for Named Entity Recognition (NER) in Catalan. It adapts <a href="http://clic.ub.edu/corpus/">AnCora corpus</a> for Machine Learning and Language Model evaluation purposes. This dataset was developed by [BSC TeMU](https://temu.bsc.es/) as part of the [Projecte AINA](https://politiquesdigitals.gencat.cat/ca/economia/catalonia-ai/aina/), to enrich the [Catalan Language Understanding Benchmark (CLUB)](https://club.aina.bsc.es/). ### Supported Tasks and Leaderboards Named Entities Recognition, Language Model ### Languages The dataset is in Catalan (`ca-ES`). ## Dataset Structure ### Data Instances Three two-column files, one for each split. <pre> Fundació B-ORG Privada I-ORG Fira I-ORG de I-ORG Manresa I-ORG ha O fet O un O balanç O de O l' O activitat O del O Palau B-LOC Firal I-LOC </pre> ### Data Fields Every file has two columns, with the word form or punctuation symbol in the first one and the corresponding IOB tag in the second one. ### Data Splits We took the original train, dev and test splits from the [UD version of the corpus](https://huggingface.co/datasets/universal_dependencies) - train: 10,630 examples - validation: 1,429 examples - test: 1,528 examples ## Dataset Creation ### Curation Rationale We created this corpus to contribute to the development of language models in Catalan, a low-resource language. ### Source Data #### Initial Data Collection and Normalization [AnCora](http://clic.ub.edu/corpus/) consists of a CatCAalan corpus (AnCora-CA) and a Spanish corpus (AnCora-ES), each of them of 500,000 tokens (some multi-word). The corpora are annotated for linguistic phenomena at different levels. AnCora corpus is mainly based on newswire texts. For more information, refer to Taulé, M., M.A. Martí, M. Recasens (2009): <a href="http://www.lrec-conf.org/proceedings/lrec2008/pdf/35_paper.pdf">"AnCora: Multilevel Annotated Corpora for Catalan and Spanish”</a>, Proceedings of 6th International Conference on language Resources and Evaluation. #### Who are the source language producers? Catalan [AnCora corpus](http://clic.ub.edu/corpus/) is compiled from articles from the following news outlets: <a href="https://www.efe.com">EFE</a>, <a href="https://www.acn.cat">ACN</a>, <a href="https://www.elperiodico.cat/ca/">El Periodico</a>. ### Annotations #### Annotation process We adapted the NER labels from [AnCora corpus](http://clic.ub.edu/corpus/) to a token-per-line, multi-column format. #### Who are the annotators? Original annotators from [AnCora corpus](http://clic.ub.edu/corpus/). ### Personal and Sensitive Information No personal or sensitive information included. ## Considerations for Using the Data ### Social Impact of Dataset We hope this corpus contributes to the development of language models in Catalan, a low-resource language. ### Discussion of Biases [N/A] ### Other Known Limitations [N/A] ## Additional Information ### Dataset Curators Text Mining Unit (TeMU) at the Barcelona Supercomputing Center (bsc-temu@bsc.es) This work was funded by the [Departament de la Vicepresidència i de Polítiques Digitals i Territori de la Generalitat de Catalunya](https://politiquesdigitals.gencat.cat/en/inici/index.html) within the framework of [Projecte AINA](https://politiquesdigitals.gencat.cat/ca/economia/catalonia-ai/aina/). ### Licensing information This work is licensed under a <a rel="license" href="https://creativecommons.org/licenses/by/4.0/">Attribution 4.0 International License</a>. [AnCora corpus](http://clic.ub.edu/corpus/) is used under [CC-by](https://creativecommons.org/licenses/by/4.0/) licence. ### Citation Information ``` @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", } ``` [](https://doi.org/10.5281/zenodo.4761746) ### Contributions [N/A]
### 数据集元数据 - 标注生成方式:专家生成 - 语料来源:采集自现有文本 - 语言:加泰罗尼亚语(ca) - 许可协议:CC BY 4.0 - 多语言属性:单语言 - 数据集名称:ancora-ca-ner - 样本规模类别:未知 - 源数据集:无 - 任务类别:无 - 任务子类型:无 # AnCora-Ca-NER 数据集卡片 ## 数据集描述 - **官网**:https://zenodo.org/records/4761746 - **主页** [Projecte AINA(阿伊纳项目)](https://projecteaina.cat/tech/) - **代码仓库** [HuggingFace](https://huggingface.co/projecte-aina) - **论文**:[Are Multilingual Models the Best Choice for Moderately Under-resourced Languages? A Comprehensive Assessment for Catalan](https://arxiv.org/abs/2107.07903) - **论文**:[AnCora: Multilevel Annotated Corpora for Catalan and Spanish](http://www.lrec-conf.org/proceedings/lrec2008/pdf/35_paper.pdf) - **联系人**:langtech@bsc.es ### 数据集概览 本数据集用于加泰罗尼亚语的命名实体识别(Named Entity Recognition, NER)任务,适配自AnCora语料库,用于机器学习与大语言模型(Large Language Model, LLM)评估。 本数据集由巴塞罗那超级计算中心文本挖掘单元(BSC TeMU)开发,作为阿伊纳项目(Projecte AINA)的一部分,旨在丰富加泰罗尼亚语言理解基准(Catalan Language Understanding Benchmark, CLUB)。 ### 支持任务与基准排行榜 命名实体识别、语言模型 ### 语言说明 本数据集采用加泰罗尼亚语(ca-ES)。 ## 数据集结构 ### 数据实例 本数据集包含三个双列文件,分别对应三个数据划分。 <pre> Fundació B-ORG Privada I-ORG Fira I-ORG de I-ORG Manresa I-ORG ha O fet O un O balanç O de O l' O activitat O del O Palau B-LOC Firal I-LOC </pre> ### 数据字段 每个文件包含两列:第一列为词形或标点符号,第二列为对应的IOB标记(IOB tag)。 ### 数据划分 我们从该语料库的通用依存(Universal Dependencies, UD)版本中获取了原始的训练、开发与测试划分: - 训练集:10,630 条样本 - 验证集:1,429 条样本 - 测试集:1,528 条样本 ## 数据集构建 ### 构建初衷 我们创建本语料库,旨在推动低资源语言加泰罗尼亚语的大语言模型发展。 ### 源数据 #### 初始数据收集与标准化 AnCora语料库包含加泰罗尼亚语语料库(AnCora-CA)与西班牙语语料库(AnCora-ES),两者均包含50万个Token(部分为多词单元),并针对多维度语言现象进行了标注。该语料库主要基于新闻文本,详细信息可参考Taulé, M., M.A. Martí, M. Recasens (2009) 发表的论文《AnCora: Multilevel Annotated Corpora for Catalan and Spanish》,收录于第6届国际语言资源与评估会议论文集。 #### 源语言生产者是谁? 本数据集所使用的加泰罗尼亚语AnCora语料库源自以下新闻媒体的文章:EFE、ACN、《El Periodico》。 ### 标注信息 #### 标注流程 我们将AnCora语料库的NER标签适配为“逐词分行、多列格式”。 #### 标注人员是谁? 标注人员为AnCora语料库的原始标注团队。 ### 个人与敏感信息 本数据集未包含任何个人或敏感信息。 ## 数据使用注意事项 ### 数据集的社会影响 我们期望本语料库能够推动低资源语言加泰罗尼亚语的大语言模型发展。 ### 偏差讨论 无相关内容(N/A) ### 其他已知局限性 无相关内容(N/A) ## 补充信息 ### 数据集维护方 巴塞罗那超级计算中心文本挖掘单元(Text Mining Unit, TeMU),联系邮箱:bsc-temu@bsc.es 本工作由加泰罗尼亚政府副主席兼数字政策与领土部依托阿伊纳项目(Projecte AINA)资助。 ### 许可协议信息 本数据集采用知识共享署名4.0国际许可协议(CC BY 4.0)进行授权。AnCora语料库采用CC BY许可协议授权。 ### 引用信息 @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-Estapé, 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", } [](https://doi.org/10.5281/zenodo.4761746) ### 贡献说明 无相关内容(N/A)
数据集概述
数据集名称
- 名称: ancora-ca-ner
- 别名: AnCora-Ca-NER
数据集描述
- 目的: 用于加泰罗尼亚语的命名实体识别(NER)。
- 开发机构: BSC TeMU(Barcelona Supercomputing Center TeMU)。
- 项目背景: 作为Projecte AINA的一部分,旨在丰富加泰罗尼亚语言理解基准(CLUB)。
语言信息
- 语言: 加泰罗尼亚语(ca-ES)。
数据集结构
- 数据实例: 包含三个两列文件,分别对应训练、验证和测试集。
- 数据字段: 每列包含词形或标点符号及其对应的IOB标签。
- 数据分割:
- 训练集: 10,630例
- 验证集: 1,429例
- 测试集: 1,528例
数据集创建
- 来源数据: 基于AnCora语料库,主要由新闻文本组成。
- 注释过程: 从AnCora语料库适应NER标签至多列格式。
- 注释者: 原始AnCora语料库的注释者。
使用注意事项
- 社会影响: 旨在促进加泰罗尼亚语这一低资源语言的语言模型发展。
许可证信息
- 许可证: CC-BY-4.0
引用信息
@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", }




