DravidianCodeMix

DravidianCodeMix是一个多语言数据集，由Insight SFI研究中心开发，专注于达罗毗荼语系中的三种语言（泰米尔语、卡纳达语、马拉雅拉姆语）与英语混合文本的情感分析和攻击性语言识别。该数据集包含超过60,000条来自社交媒体的评论，这些评论经过志愿者标注，具有高的一致性。数据集不仅涵盖了各种代码混合现象，还为每种语言提供了详细的情感和攻击性标签，适用于多种自然语言处理任务，特别是在处理低资源语言和社交媒体内容分析方面具有重要价值。

DravidianCodeMix is a multilingual dataset developed by the Insight SFI Research Centre, focusing on sentiment analysis and offensive language recognition for code-mixed texts combining three Dravidian languages (Tamil, Kannada, Malayalam) with English. This dataset contains over 60,000 social media comments, which were annotated by volunteers with high annotation consistency. It covers various code-mixing phenomena, and provides detailed sentiment and offensive labels for each target language, making it applicable to a wide range of natural language processing tasks. It is particularly valuable for low-resource language processing and social media content analysis.