tweetfeels-100k

# TweetFeels 100k

A compact, automatically labeled Twitter-sentiment corpus of 101 160 tweets collected via the Twitter Search API in 2015. Polarity labels (0 = negative, 4 = positive) were assigned using distant supervision: tweets containing positive emoticons such as “:)” were deemed positive, while those with negative emoticons like “:(” were deemed negative.

Each record carries two tab-separated fields:

- **target** – sentiment label (0 or 4)
- **tweet** – full UTF-8 tweet text

The dataset focuses on short, informal language typical of Twitter and is sized for rapid experimentation, classroom use, or quick model prototyping in sentiment-analysis pipelines.

**Acknowledgements**: The dataset is hosted on [Kaggle—Tweet Sentiment Dataset](https://www.kaggle.com/datasets/farheenshaukat/tweet-sentiment-dataset).

# TweetFeels 100k

TweetFeels 100k：一款紧凑的自动标注推特情感语料库，于2015年通过推特搜索API（Twitter Search API）采集了101160条推文。该数据集的情感极性标签（0代表负面情感，4代表正面情感）采用远程监督（distant supervision）方式标注：包含“:)”等正面表情符号的推文被标记为正面样本，包含“:(”等负面表情符号的推文则被标记为负面样本。

每条数据包含两个以制表符分隔的字段：
- **target**：情感极性标签（取值为0或4）
- **tweet**：完整UTF-8编码的推文内容

本数据集聚焦推特平台典型的简短非正式语言，体量适配快速实验、课堂教学或情感分析流水线中的快速模型原型构建需求。

**致谢**：本数据集托管于[Kaggle—推特情感数据集（Tweet Sentiment Dataset）](https://www.kaggle.com/datasets/farheenshaukat/tweet-sentiment-dataset).