基于加权二部图的课程教师推荐方法  

Recommendation Method of the Course Teachers Based on Weighted Bipartite Graph

作　　者：姚敦红[1] 邓小武[1] Yao Dunhong;Deng Xiaowu(School of Computer Science and Artificial Intelligence,Huaihua University,Huaihua 418000,China)

机构地区：[1]怀化学院计算机与人工智能学院,怀化418000

出　　处：《办公自动化》2023年第16期5-8,48,共5页Office Informatization

基　　金：湖南省社会科学成果评审委员会课题(项目编号:XSP21YBC359)。

摘　　要：为给高校课程推荐合适的授课教师以提高课程教学质量,本研究提出一种基于加权二部图的课程授课教师推荐方法。首先,采用TSTI量表和TPACK测量工具,以及所构建的教学质量、课程难度等量表,收集某学科专业中的教师和课程特征数据,并构建教师-课程-教学质量稀疏实验矩阵。接着,将课程难度和教学质量特征混入二部图,构建加权二部图模型,并采用改进的PersonalRank算法科学地预测出教师在未授课程上的教学质量。最后,创建一种融合教师的教学风格和TPACK特征的TOP-N推荐模型,实现课程的授课教师的精确推荐。对比实验结果显示,该方法在课程授课教师推荐准确性和性能上优于传统的协同过滤(CF)推荐算法,并且该方法具有可解释性。To recommend the suitable teachers for university courses and improve the quality of course instruction,this study proposes a course teachers recommendation method based on a weighted bipartite graph.Firstly,the TSTI scale table,TPACK measurement tool,and teaching quality and course difficulty scales were constructed to collect ef-fectively the teachers and course feature data in a certain subject major and construct a sparse experimental matrix of teacher-course-teaching quality.Then,the course difficulty and teaching quality features were mixed into the bipartite graph to construct a weighted bipartite graph model,and the improved PersonalRank algorithm was used to predict the teachers'teaching quality on untaught courses scientifically.Finally,a TOP-N recommendation model that inte-grates the teacher's teaching style and TPACK features was created to recommend course instructors accurately.Com-parative experimental results show that this method is superior to traditional collaborative filtering(CF)recommenda-tion algorithms in terms of accuracy and performance of the course teachers recommendation and possesses this method interpretability.

关 键 词：教学风格 技术的学科教学知识(TPACK) 课程教学质量 加权二部图 推荐模型 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术] G642.3[自动化与计算机技术—计算机科学与技术] G451[文化科学—高等教育学]