google/civil_comments
收藏资源简介:
该数据集包含来自Civil Comments平台的公开评论,该平台是一个独立新闻网站的评论插件。这些评论创建于2015年至2017年之间,出现在全球大约50个英文新闻网站上。当Civil Comments在2017年关闭时,他们选择将这些公开评论保存在一个持久的开放档案中,以便未来的研究使用。原始数据包括公开评论文本、一些相关的元数据(如文章ID、时间戳和评论者生成的“文明”标签),但不包括用户ID。Jigsaw通过添加额外的毒性标签和身份提及标签扩展了这个数据集。该数据集是Jigsaw在Kaggle上发布的“毒性分类中的无意偏见”挑战中使用的数据的精确复制品。该数据集及其基础评论文本均以CC0许可证发布。
This dataset contains public comments sourced from the Civil Comments platform, a comment plugin for independent news websites. These comments were created between 2015 and 2017, and appeared on approximately 50 English-language news websites worldwide. When Civil Comments shut down in 2017, they opted to preserve these public comments in a permanent open archive for future research use. The original dataset includes public comment text, associated metadata such as article IDs, timestamps, and commenter-generated "civil" labels, but excludes user IDs. Jigsaw expanded this dataset by adding additional toxicity labels and identity mention labels. This dataset is an exact replica of the data used in Jigsaw's "Unintended Bias in Toxicity Classification" challenge hosted on Kaggle. Both this dataset and the underlying comment text are released under the CC0 license.
数据集概述
数据集基本信息
- 名称: Civil Comments
- 许可证: CC0-1.0
- 标签: toxic-comment-classification
- 任务类别: text-classification
- 任务ID: multi-label-classification
数据集结构
数据特征
- text: 字符串类型
- toxicity: 浮点数类型 (float32)
- severe_toxicity: 浮点数类型 (float32)
- obscene: 浮点数类型 (float32)
- threat: 浮点数类型 (float32)
- insult: 浮点数类型 (float32)
- identity_attack: 浮点数类型 (float32)
- sexual_explicit: 浮点数类型 (float32)
数据分割
- 训练集: 1804874 样本
- 验证集: 97320 样本
- 测试集: 97320 样本
数据集创建
许可证信息
- 许可证: CC0 1.0
引用信息
@article{DBLP:journals/corr/abs-1903-04561, author = {Daniel Borkan and Lucas Dixon and Jeffrey Sorensen and Nithum Thain and Lucy Vasserman}, title = {Nuanced Metrics for Measuring Unintended Bias with Real Data for Text Classification}, journal = {CoRR}, volume = {abs/1903.04561}, year = {2019}, url = {http://arxiv.org/abs/1903.04561}, archivePrefix = {arXiv}, eprint = {1903.04561}, timestamp = {Sun, 31 Mar 2019 19:01:24 +0200}, biburl = {https://dblp.org/rec/bib/journals/corr/abs-1903-04561}, bibsource = {dblp computer science bibliography, https://dblp.org} }




