CompRes
收藏arXiv2023-11-07 更新2024-08-06 收录
下载链接：
http://arxiv.org/abs/2007.04874v2
下载链接
链接失效反馈官方服务：
资源简介：
CompRes数据集是由希伯来大学创建的，专注于新闻媒体中的叙事结构分析。该数据集包含29篇英文新闻文章，总计1099个句子，每篇文章来自不同的主流和党派网站。数据集通过设计新的叙事标注方案，适应新闻媒体的特殊需求，包括从Labov和Waletzky的叙事理论中改编的复杂性和解决元素，以及新增的成功元素。该数据集旨在通过机器学习模型自动识别新闻文本中的叙事元素，以理解和分析新闻如何塑造公众意见和影响社会。
The CompRes Dataset was developed by The Hebrew University of Jerusalem, focusing on narrative structure analysis in news media. This dataset comprises 29 English news articles, totaling 1099 sentences, with each article sourced from distinct mainstream and partisan news websites. Designed with a novel narrative annotation scheme tailored to the specific needs of news media, the dataset incorporates complexity and resolution elements adapted from the narrative theory proposed by Labov and Waletzky, as well as a newly added success element. This dataset aims to automatically identify narrative elements in news texts via machine learning models, thereby facilitating the understanding and analysis of how news shapes public opinion and influences society.
提供机构：
希伯来大学
创建时间：
2020-07-09
