基于机器学习的高校学生评教信度分类分析

The Classification of the Reliability of College Students'Evaluation of Teaching Based on Machine Learning

大学生在评教过程中存在态度敷衍、不客观的现象,如何使学生评教更加客观,改善学生评教质量是当前教师评教过程中亟待解决的问题。目前研究主要是对学生评教数据中偏离数据的分析、处理,在一定程度上解决了学生评教不客观的问题。该研究旨在通过对学生评教数据进行分析,多角度提取反映学生客观评教信度的指标,并对指标进行相关性分析,剔除冗余指标,最后利用K-Means++算法对学生评教信度分类,确定大学生在评教这一活动中具有的群体特征。对不同群体类别的学生评教信度进行星级反馈,起到警示、敦促学生的作用,进而提高学生评教信度,使教师评价更加公平、有效。

Perfunctory and non-objective attitude of college students can be seen in the evaluation process of teaching.How to make teaching evaluation more objective and improve the quality of it has become urgent issues.The current research is mainly on the analysis and processing of deviation data in evaluation data,which only solves the problem of non-objective evaluation of students to a certain extent.The purpose of this study is to analyze the students'evaluation data and extract more indicators which can objectively reflect students'evaluation reliability.Then,redundant indicators can be removed based on their correlations.Finally,K-means++algorithm is proposed to classify the teaching evaluation reliability.The experiment results show that college students have group characteristics in the teaching evaluation activities.Different categories of students are identified as different star levels,which can serve as feedback on the teaching evaluation reliability.By warning or urging students,the method can improve the credibility of teaching evaluation,and it can also make the results of evaluation fairer and more effective.