基于CNN的教学质量评估模型研究  

作　　者：孟庆祥[1] 王帆 冯苑君 申力[1] 卢冰 杜娟[1] 

机构地区：[1]武汉大学遥感信息工程学院,武汉430000

出　　处：《高教学刊》2023年第S01期14-17,共4页Journal of Higher Education

基　　金：湖北省科技厅省级基金项目“基于全天空成像仪和辐射传输模型的太阳能短时预测方法研究”(2019CFB732);武汉大学遥感信息工程学院“三全育人”教改项目“疫情期间遥感大类卓越工程师人才培养改革”(YGJY202210);武汉大学遥感信息工程学院“三全育人”教改项目“探索‘教与学革命’背景下教师教学能力提升优化及制度设计”(YGJY202214);武汉大学遥感信息工程学院“三全育人”教改项目“价值引领式的大学生竞赛科研创新育人模式改革路径探索”(YGJY202114);武汉大学教改项目“学科竞赛与大创科研驱动的遥感类综合创新育人模式研究”(00030791)。

摘　　要：高校教学质量评估活动开展为广大教师及时调整教学策略提供针对性建议,有助于提高教学质量。但现行评估方法多为简单加权平均,人为给定指标权重,常存在主观性和片面性局限。针对上述现象,该文综合学生、同行教师、督导员三方评价将教学质量分为优秀、良好、一般和较差4个等级,并采用神经网络算法,建立基于卷积神经网络的教学质量评估模型。该文首先研究现行教学质量评估体系优缺点,判断模型构建可行性,提出基于深度学习的教学质量评估模型构建方法,然后代入现有教学评估数据,比较所得结果与已知事实验证模型准确性。结果证明,评估模型准确性较高。该研究将有助于建立一个更加科学可靠的教学质量评估标准体系。The development of teaching quality assessment activities in colleges and universities can provide more targeted information for teachers to adjust teaching strategies in a timely manner and do college lecturers a favor to improve teaching quality.Nevertheless,the majority of the current evaluation methods are weighted average algorithms.The weights of each indicator are artificially given empirical values,which often have limitations of subjectivity and one-sidedness.In view of the above phenomenon,this paper not only comprehensively considers the course evaluations of students,peer teachers and supervisors to divide the teaching quality into four grades:excellent,good,average and inferior,but also uses neural network algorithms to establish a teaching quality evaluation model structure based on Convolutional Neural Networks(CNN).This paper studies the pros and cons of the current teaching quality evaluation system,judges the feasibility of building a CNN model,and proposes a construction method for the model of teaching quality evaluation based on deep learning.Subsequently,the data of existing teaching evaluation is substituted into the model for calculation and the obtained results are compared with known facts to verify the accuracy of the model.According to the result,the accuracy of the CNN evaluation model is high,which will help to establish a more scientific and reliable teaching quality evaluation standard system.

关 键 词：教学质量 深度学习 评估方法 CNN网络 评价体系 

分 类 号：G642[文化科学—高等教育学]