FOOTBALL
收藏arXiv2019-10-19 更新2024-06-21 收录
下载链接：
http://github.com/jmerullo/football
下载链接
链接失效反馈官方服务：
资源简介：
数据集FOOTBALL由麻省大学阿默斯特分校创建，包含1455场跨越六个十年的美式足球比赛广播转录，自动标注了约25万次球员提及，并关联了种族元数据。数据集内容丰富，涵盖了球员的种族和位置信息，通过从YouTube收集的广播转录和手动标注的方式创建。该数据集主要用于研究体育评论中的种族偏见问题，旨在通过大规模计算分析支持社会科学研究中的结论。
The FOOTBALL dataset was developed by the University of Massachusetts Amherst. It contains 1,455 American football game broadcast transcripts spanning six decades, with approximately 250,000 automatically annotated player mentions paired with racial metadata. The dataset covers comprehensive information including players’ racial identities and on-field positions, and was constructed using broadcast transcripts collected from YouTube alongside manual annotations. This dataset is primarily intended for research on racial bias in sports commentary, aiming to support research conclusions in social science through large-scale computational analysis.
提供机构：
麻省大学阿默斯特分校
创建时间：
2019-09-08
