机器学习在高校课程教学评价中的应用研究

Research on the Application of Machine Learning in Teaching Evaluation for College Courses

基于学生感知的课程在线评价是评价教师绩效的常用工具之一。本文利用决策树算法、支持向量机、贝叶斯和随机森林四种不同的分类技术,对学生课程评价数据集进行分类实验。结果显示,四个分类器模型的分类性能都较高,其中随机森林分类器在准确率、查准率和查全率和F1值上都是最好的。此外,还分析了每个分类器模型中各个评价特征的重要性,可以用于进一步完善评价指标体系。实验结果表明,学生对课程的评价主要取决于他们对课程的兴趣。本文验证了机器学习模型在课程评价和高等教育挖掘中的有效性。

Online course evaluation based on students’perception is one of the common tools to evaluate teacher’s performance.In this paper,four different classification technologies,including Decision Tree algorithm,Support Vector Machine,Bayes,and Random Forest,are used to classify the data set of students’curriculum evaluation.The results show that all these four classifier models have high classification performance,among them the random forest classifier is the best in terms of accuracy,precision,recall,and F1 value.In addition,the importance of each evaluation feature for each classifier model is analyzed to further improve the evaluation index system.The result of the analysis shows that students’evaluation on courses mainly depends on students’interest in courses.This research verifies the validity of machine learning model in curriculum evaluation and higher education mining.