MOOPer

该数据集是一个广泛的在线开放实践数据集，涵盖了2018年至2019年间学生的在线练习数据，同时还包括了以知识图谱形式组织的用户及练习项目补充信息。数据集包含11个独特的实体类别和十种具体的关系类型，为实体之间提供了分层的结构。规模上，数据集包含了约253万条练习记录，2160万条系统反馈记录以及1.5万条论坛聊天记录。任务方面，旨在实现MOOCs中的自动评分预测和早期退学预测。This dataset is a comprehensive online open practice dataset, covering students' online practice data from 2018 to 2019, as well as supplementary information about users and practice projects organized in the form of knowledge graphs. The dataset includes 11 unique entity categories and ten specific relationship types, which provides a hierarchical structure between entities. In terms of scale, the dataset contains approximately 2.53 million practice records, 21.6 million system feedback records, and 15,000 forum chat records. For downstream tasks, this dataset is designed for automatic scoring prediction and early dropout prediction in MOOCs.

搜集汇总

背景与挑战

背景概述

MOOPer数据集是一个用于MOOC成绩预测的数据集，包含学生与挑战互动的详细记录，并通过图表示学习技术进行增强以提高预测准确性。数据集格式为每行代表一个学生-挑战互动，包含多种特征如用户ID、挑战ID、时间戳、课程ID、难度、重试状态、持续时间和最终分数。