VSoLSCSum-Dataset
收藏资源简介:
该数据集包含141篇开放领域的文章,共3,760个句子,2,448个提取的标准句子和评论作为标准摘要,以及6,926个评论分布在12个事件中。数据集由人工手动标注,标准摘要中也包含评论。句子和评论的标签是通过社会标注者投票生成的,标签范围从1到5,表示从非常差到完美。标准摘要句子是那些至少获得三个标注者同意的句子。标注者之间的协议通过Cohens Kappa计算,验证后的值为0.685。
This dataset comprises 141 open-domain articles, totaling 3,760 sentences, 2,448 extracted standard sentences and reviews as standard summaries, and 6,926 reviews distributed across 12 events. The dataset was manually annotated by humans, and the standard summaries also include reviews. The labels for sentences and reviews were generated through voting by social annotators, with labels ranging from 1 to 5, indicating from very poor to perfect. Standard summary sentences are those that have received agreement from at least three annotators. The agreement among annotators was calculated using Cohen's Kappa, with a validated value of 0.685.
数据集概述
数据集名称
VSoLSCSum-Dataset
数据集用途
用于越南语社交情境摘要
数据集内容
- 包含141篇开放领域的文章
- 总计3,760个句子
- 提取的标准句子及评论共2,448条,作为标准摘要
- 包含6,926条评论,分布在12个事件中
数据集特点
- 数据集由人工手动标注
- 每个句子的标签通过社交标注者的投票产生,标签范围从1到5,代表不同的质量等级
- 标准摘要句子需至少获得三位标注者的同意
- 标注者之间的协议度通过Cohens Kappa计算,验证后的值为0.685




