BasqueParl

BasqueParl是一个新编译的双语语料库，源自巴斯克议会的转录文本，涵盖2012至2020年的议会辩论记录。该数据集主要特点是巴斯克语和西班牙语之间的频繁代码转换，为研究政治话语提供了独特资源。数据集包含超过1400万单词，通过添加与演讲者和演讲相关的元数据（如语言、性别、政党等）进行了丰富，并进行了词形还原和命名实体识别处理。该数据集旨在通过多语言和跨语言研究，解决与论证、话语结构、情感分析和事实核查相关的NLP任务。BasqueParl is a newly compiled bilingual corpus derived from the transcribed texts of the Basque Parliament, covering parliamentary debate records from 2012 to 2020. Its core feature is frequent code-switching between Basque and Spanish, providing a unique resource for political discourse research. The corpus contains over 14 million words, has been enriched with metadata associated with speakers and their speeches (such as language, gender, political party, etc.), and has been processed with lemmatization and named entity recognition. This dataset aims to address NLP tasks related to argumentation, discourse structure, sentiment analysis and fact-checking through multilingual and cross-linguistic research.