awaisafridi55/argument_quality_ranking_30k

该数据集包含30,497个众包论点，涉及71个争议性话题，每个论点都有质量和立场标签，分为训练集、验证集和测试集。Argument Topic子集包含9,487个论点，仅包含话题信息。数据集通过众包任务收集，要求参与者为每个话题提供一个支持和反对的论点，并对论点的质量和立场进行标注。每个实例包含论点文本、话题、质量评分（WA和MACE-P）和立场评分（stance_WA和stance_WA_conf）。

The dataset contains 30,497 crowd-sourced arguments for 71 debatable topics labeled for quality and stance, split into train, validation and test sets. The Argument Topic subset contains 9,487 of the arguments only with their topics. The dataset was collected through a crowd annotation task, where participants were asked to contribute one supporting and one contesting argument for each topic, and label the quality and stance of the arguments. Each instance contains the argument text, topic, quality scores (WA and MACE-P), and stance scores (stance_WA and stance_WA_conf).