面向雨课堂的数据分析与应用研究

Empirical Study of Data Mining Based on Rain Classroom

以学生在雨课堂中积累的课程数据为基础,从学习态度、学习能力两个方面获得阅读率、提交率、到课率、得分率等多个维度的特征向量,并进行数据挖掘实证研究。首先,通过聚类对全体学生每周的学习状态进行分析,并将每个学生在学习进程中的聚类变化进行可视化;其次,通过聚类将学生分成“优秀”“良好”“中等及以下”三类,分别对三类学生进行不同特征的相关分析;最后,利用机器学习的4种模型对课程数据进行预测。结果表明,学习态度与学习能力表现出高度相关性,其中作业得分率、作业提交率、课堂得分率的相关性排在前三位;预测模型能够对学生未来的课程学习效果进行评估。

Based on the course data accumulated by the students in the Rain Classroom,the read rate,submitting rate,attendance rate,scoring rate,and other feature vectors are obtained from the two aspects of learning attitude and learning ability.According to these data,clustering analyze can help us to grasp the students'learning status firstly.Seconclly,students can be clustered into three groups,excellent,good and others.We utilize the correlation analyze to varied features among students.Thirdly four machine learning models are used to predict the students'learning effects.The results show that learning effect is closely related with scoring rate of assignment,and submitting rate of assignment.The features and models can be used to predict students'future learning effects with good accuracy.