基于文本挖掘的教学质量评价指标量化研究

Quantitative Research on Teaching Quality Evaluation Index Based on Text Mining

摘要: 教学质量评价数据主要由专家给出的评判等第和综合评语构成，现存的教学质量评估方法只能根据专家给出的等第对教学质量进行粗略划分，缺少对于专家综合评语的具体量化。本文旨在通过文本挖掘方法实现对于专家评语的具体量化，然后结合专家给出的等第进行监督加权，得到一个结合专家等地和专家评分的综合量化得分，实现对于教学质量的具体评价。最后使用提出的量化指标对该校近三年的教学质量进行稀疏主成分分析，得到的三个主成分具有良好的解释性，可以被解释为“学期效应”、“理论课效应”和“实践课效应”，揭示出新冠疫情爆发导致高校教学质量急剧降低的规律。

Abstract: The teaching quality evaluation data are mainly composed of the evaluation grade and comprehen-sive comments given by experts. The existing teaching quality evaluation methods can only roughly divide the teaching quality according to the grade given by experts, and lack of specific quantifica-tion of experts’ comprehensive comments. This paper aims to achieve the specific quantification of expert comments through text mining method, and then combine the grade given by experts to su-pervise and weight to obtain a comprehensive quantitative score combining expert grade, so as to achieve the specific evaluation of teaching quality. Finally, the proposed quantitative indicators are used to conduct sparse principal component analysis on the teaching quality of the school in the past three years, and the three principal components obtained have good explanatory power, which can be interpreted as “semester effect”, “theory course effect” and “practice course effect”, revealing the law that the outbreak of the COVID-19 has led to a sharp decline in the teaching quality of col-leges and universities.