tweetfeels-1m4
收藏魔搭社区2025-12-05 更新2025-12-06 收录
下载链接：
https://modelscope.cn/datasets/mnemoraorg/tweetfeels-1m4
下载链接
链接失效反馈官方服务：
资源简介：
# TweetFeels 1m4
An 1-million-tweet sentiment corpus harvested from Twitter and annotated with four fine-grained categories: positive, negative, uncertainty, and litigious. Each record carries three clean, tab-separated fields:
- **Language** – ISO-639 code of the tweet’s detected language
- **Text** – the full tweet text, preserved with original casing and emojis
- **Label** – one of {positive, negative, uncertainty, litigious} determined by an automated labelling pipeline
The dataset covers multiple languages and informal, real-time expressions typical of Twitter, offering a sizeable, ready-to-use resource for multi-class sentiment and legal-tone modelling.
**Acknowledgements**: The dataset is hosted on [Kaggle—Sentiment Dataset with 1 Million Tweets](https://www.kaggle.com/datasets/tariqsays/sentiment-dataset-with-1-million-tweets)
# TweetFeels 1m4
本数据集为从推特（Twitter）采集的百万级推文情感语料库，共标注四类细粒度情感类别：积极、消极、不确定及诉讼相关。每条数据均包含三个经清洗的制表符分隔字段：
- **语言** – 推文检测语言对应的ISO-639标准代码
- **文本** – 完整推文原文，保留原始大小写格式与表情符号
- **标签** – 属于{积极、消极、不确定、诉讼相关}四类之一，由自动化标注流水线生成
该数据集涵盖多语种内容，包含推特平台典型的非正式实时表达，可为多分类情感建模与法律语气建模提供规模可观的现成可用资源。
**致谢**：本数据集托管于[Kaggle——百万推文情感数据集](https://www.kaggle.com/datasets/tariqsays/sentiment-dataset-with-1-million-tweets)
提供机构：
maas
创建时间：
2025-09-08
