Student Evaluations of Generated Questions
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/limu-research/2024-ectel-qg
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了246个不同问题，由学生们在五个维度上进行评价，分别是清晰度、相关性、难度、幻灯片关联性和问题答案一致性，总共产生了5339个评估结果。此外，数据集还包含了基于信心水平的评估，并使用了Kruskal-Wallis H检验进行了大量分析。规模上，有46名学生对246个问题进行了评价。该数据集的任务是对自动问题生成评估进行研究。
This dataset contains 246 distinct questions, which were evaluated by students across five dimensions: clarity, relevance, difficulty, slide relevance, and question-answer consistency, generating a total of 5339 evaluation results. Additionally, the dataset includes evaluations based on confidence levels, and extensive analyses were conducted using the Kruskal-Wallis H test. In terms of scale, 46 students evaluated the 246 questions. The core task of this dataset is to support research on automatic question generation evaluation.
提供机构：
University students
