Bhojpuri, Maithili, Magahi NER Dataset

Bhojpuri, Maithili, Magahi NER Dataset是由印度理工学院(BHU)计算机科学与工程系的研究团队为三种低资源语言Bhojpuri, Maithili, Magahi开发的命名实体识别基准数据集。该数据集旨在支持从这些语言到印地语的机器翻译系统的开发。数据集包含三个子集，分别对应每种语言，总计标注了22种实体标签，包括人名、地点、组织等。Bhojpuri子集包含228373个令牌，Maithili子集157468个令牌，Magahi子集56190个令牌。这些数据集通过精细的标注过程，为每种语言提供了丰富的实体识别训练资源，特别适用于解决低资源语言在机器翻译中的实体识别问题。

The Bhojpuri, Maithili, Magahi NER Dataset is a named entity recognition (NER) benchmark dataset developed by a research team from the Department of Computer Science and Engineering, Indian Institute of Technology (BHU) for three low-resource languages: Bhojpuri, Maithili, and Magahi. This dataset is designed to support the development of machine translation systems that translate text from these languages into Hindi. The dataset includes three subsets corresponding to each of the three languages, with a total of 22 annotated entity tags covering personal names, locations, organizations, and other common entity categories. The Bhojpuri subset contains 228,373 tokens, the Maithili subset contains 157,468 tokens, and the Magahi subset contains 56,190 tokens. Developed through rigorous annotation procedures, these datasets provide abundant training resources for entity recognition tasks for each language, and are particularly suitable for addressing entity recognition challenges faced by low-resource languages in machine translation.