Corpus of Anglicisms in the Spanish Press (COALAS)

Corpus of Anglicisms in the Spanish Press (COALAS) 是一个专门用于研究西班牙新闻中未同化英语借词的标注语料库。该数据集由国立远程教育大学的计算机科学与工程学院创建，包含372,701个令牌，主要来源于西班牙的在线新闻文章。数据集旨在通过CRF、BiLSTM-CRF和Transformer等序列标注模型，评估和分析在识别借词任务上的性能。数据集不仅覆盖了广泛的主题，还特别关注了借词密集和OOV（Out-Of-Vocabulary）丰富的文本，以更好地评估模型的泛化能力。COALAS数据集的应用领域包括语言学研究、自然语言处理以及借词自动检测等，旨在解决借词识别和语言变化分析中的问题。

Corpus of Anglicisms in the Spanish Press (COALAS) is an annotated corpus specifically dedicated to researching non-assimilated English loanwords in Spanish press content. This dataset was developed by the School of Computer Science and Engineering at the National University of Distance Education, and it contains 372,701 tokens derived primarily from Spanish online news articles. The dataset aims to evaluate and analyze the performance of sequence labeling models such as CRF, BiLSTM-CRF, and Transformer on loanword identification tasks. Beyond covering a wide range of topics, the corpus also places special focus on text that is dense with loanwords and rich in out-of-vocabulary (OOV) terms, so as to better assess the generalization ability of models. Application areas of the COALAS corpus include linguistic research, natural language processing, automatic loanword detection and other related fields, with the goal of addressing challenges in loanword identification and language change analysis.