x-stance

x-stance 数据集包含 150 多个政治问题，以及候选人就这些问题撰写的 67k 条评论。它可用于训练和评估姿态检测系统。评论部分是德语，部分是法语和意大利语。这些问题以所有三种语言和英语提供。来自瑞士选举候选人所写评论的大规模姿态检测数据集。该数据集由德语、法语和意大利语文本组成，允许对姿态检测进行跨语言评估。它包含对 150 多个政治问题（目标）的 67 000 条评论。The x-stance dataset is a large-scale stance detection dataset consisting of comments written by Swiss election candidates. It contains 67,000 comments on over 150 political topics (targets), all authored by candidates regarding these issues. Most of the comments are in German, with some samples in French and Italian. All topics are available in German, French, Italian as well as English. The dataset, which comprises text in German, French and Italian, enables cross-language evaluation of stance detection systems and can be utilized for training and evaluating such systems.

搜集汇总

背景与挑战

背景概述

x-stance是一个用于姿态检测的多语言数据集，包含超过150个政治问题和67,000条评论，覆盖德语、法语、意大利语和英语。该数据集由爱丁堡大学和苏黎世大学于2020年发布，适用于训练和评估跨语言姿态检测系统。