NagaNLP POS Annotated Corpus

The NagaNLP POS Annotated Corpus is a human-validated Part-of-Speech annotated dataset for Nagamese, containing 214 sentences and approximately 4,800 tokens. Each token is assigned a Universal POS tag to support POS tagging and low-resource NLP research. This dataset is a part of NagaNLP Toolkit.

About Dataset

The NagaNLP POS Annotated Corpus is a Part-of-Speech annotated resource developed for Nagamese (Naga Pidgin), a low-resource creole widely used as a lingua franca across Nagaland and the wider Northeast India region. The dataset contains 214 sentences and approximately 4,800 tokens. Each instance consists of a tokenized sentence and its corresponding Universal POS tags, including categories such as NOUN, VERB, PROPN, ADJ, ADV, PRON, ADP, SCONJ, CCONJ, PART, NUM, and PUNCT. The sentences cover regional news and everyday topics including weather, sports, local government, and current affairs. The dataset shares its underlying sentences with the NagaNLP NER Annotated Corpus but is provided as a standalone resource for users requiring POS annotations. The corpus was created as part of the NagaNLP project. The dataset is described in: Agniva Maiti, Manya Pandey, and Murari Mandal. (2025). NagaNLP: Bootstrapping NLP for Low-Resource Nagamese Creole with Human-in-the-Loop Synthetic Data. arXiv preprint arXiv:2512.12537. This dataset was identified and facilitated for onboarding as part of the Dataset Onboarding Support Team (DOST) initiative led by CivicDataLab (CDL), partnering with the Gates Foundation in collaboration with BHASHINI. CivicDataLab provided support for dataset discovery, validation, metadata preparation, and onboarding facilitation. All dataset ownership and intellectual property rights remain with the original author(s).

Purpose of Dataset

The purpose of the dataset is to provide a foundational POS-tagged resource for Nagamese and support the development and evaluation of NLP systems for this low-resource language. It can be used for Part-of-Speech tagging, token classification, linguistic analysis, syntactic processing, language model development, low-resource NLP research, and evaluation of multilingual and cross-lingual NLP systems. It can also support language technology development and preservation for Nagamese.

Dataset Metadata

License

Attribution-Non-Commercial 4.0 International (CC BY-NC 4.0)

Geographical coverage

Northeast India

Sector

Science, Technology and Research

Author

Agniva Maiti

Source Organisation

Digital India BHASHINI Division

Uploaded by

Nikil Augustine

Data Quality Score (Beta)

Dataset type

Structured

Frequency

Static

Time Granularity

Static

Year range

N.A.

Date & Time

26/08/26 07:53:57

Visibility

Open

Hosted / Redirected

Hosted

Data Type

Hybrid

Data Collection Method

This Dataset Is A Part Of Naganlp Toolkit. The Corpus Was Created Through An Llm-to-human Bootstrapping Pipeline. Gemini 2.5 Pro Was Used For Interactive Nagamese Grammatical Elicitation, Knowledge Consolidation, And Few-shot Generation Of Synthetic Sentences Using Authentic Nagamese Texts As Reference. Native Speakers Then Corrected The Generated Sentences For Grammatical Accuracy. A Nagamese-aware Llm Performed Initial Pos Annotation, Which Was Subsequently Reviewed And Corrected By Human Annotators To Produce The Gold-standard Corpus. Pos Annotation Follows The Universal Dependencies (Ud) V2 Tagset, With Code-switched English Words Tagged According To Their Grammatical Function. Annotation Quality Was Validated Through Independent Human Annotation, Achieving Cohen's Kappa Of 0.92 For Pos.