ELNER-DZ: A Dataset for Named Entity Recognition and Entity Linking in Algerian Arabic Dialect

ELNER-DZ is the first large-scale dataset for Named Entity Recognition (NER) and Entity Linking (EL) in Algerian Arabic (Darija), including Arabic script and Arabizi (Latin-script). It was developed as part of a Master’s thesis . The dataset contains over 2 million dialectal sentences labeled with more than 1.9 million named entities linked to Wikidata QIDs. It includes annotations in JSON format and is suitable for NLP, NER, and EL tasks on low-resource, dialectal, and code-switched data.ELNER-DZ是首个针对阿尔及利亚阿拉伯语（达里贾语，Darija）的命名实体识别（Named Entity Recognition，NER）与实体链接（Entity Linking，EL）大规模数据集，同时涵盖阿拉伯文书写体与阿拉伯拉丁化书写体（Arabizi）。该数据集作为硕士学位论文的一部分开发完成。 该数据集包含超200万条方言语句，标注有超过190万个关联至Wikidata QID的命名实体。其标注文件采用JSON格式存储，可适用于低资源、方言类及语码转换数据场景下的自然语言处理（NLP）、命名实体识别与实体链接任务。