Local Indonesian NLP dataset

本研究旨在创建一个针对印尼地方语言的低资源语言数据集，特别是针对印尼地方语言的自然语言处理（NLP）数据集。数据集的构建面临多种挑战，包括语言多样性、地理和文化影响。研究使用Twitter的地理位置工具进行自动标注，通过过滤广泛使用的外国语言推文，并区分印尼文本的正式与非正式形式，最终尝试将推文分类到其发布地区。该数据集的应用领域包括提升对印尼地方语言的理解和处理能力，解决语言多样性和标准化缺乏的问题。

This study aims to develop a low-resource language dataset targeting Indonesian regional languages, specifically a natural language processing (NLP) dataset tailored for these linguistic varieties. Multiple challenges arise during the dataset construction process, including linguistic diversity, as well as geographical and cultural influences. This work employs Twitter's geolocation tools to perform automatic annotation, filters out widely used foreign-language tweets, distinguishes between formal and informal variants of Indonesian text, and ultimately classifies tweets into their corresponding originating regions. The application domains of this dataset include enhancing the understanding and processing capabilities of Indonesian regional languages, as well as addressing the challenges posed by linguistic diversity and the lack of standardization for these local language varieties.