Multilingual Tweets from 32 Countries
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/ResponsibleAILab/emoji-universal-sentiment
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了与2018年FIFA足球世界杯相关的推文，这些推文来自32个参赛国家和地区。数据集的特色在于使用了GPT-4o生成的情感标签。此外，该数据集不仅包含了原始的多语种推文，还包括了它们的英文翻译以及相应的情感标签，涵盖19种语言。规模上，该数据集分为32个子集，每个国家或地区包含8万条推文。这项任务是对表情符号进行情感分析。
This dataset comprises Tweets related to the 2018 FIFA World Cup, collected from 32 participating countries and regions. A key highlight of this dataset is the inclusion of sentiment labels generated via GPT-4o. In addition to the original multilingual Tweets, the dataset provides their English translations and corresponding sentiment labels, covering a total of 19 distinct languages. Regarding its scale, the dataset is split into 32 subsets, with each subset containing 80,000 Tweets corresponding to one participating country or region. The downstream task associated with this dataset is emoji-based sentiment analysis.
提供机构：
Collected via Twitter Streaming APIs
