基于MLP和RBF神经网络模型预测在线开放课程学习者满意度  

Online open course learner satisfaction prediction based on MLP and RBF neural network models

作　　者：赵涵 李康 ZHAO Han;LI Kang(Bengbu Campus of Army Academy of Armored Forces,Bengbu 233000,China)

机构地区：[1]陆军装甲兵学院蚌埠校区,安徽蚌埠233000

出　　处：《现代电子技术》2023年第17期84-88,共5页Modern Electronics Technique

基　　金：高等教育科学研究项目(2022CG75)。

摘　　要：网络课程在近10年的发展中,在线学习者的满意度一直是一个重要的理论热点。依托某大规模在线教学与学习平台的虚拟学习环境,基于某校在网络学习空间开设的9门在线开放课程,结合LMS系统中的在线学习行为分析数据,研究开发、训练了具有不同激活函数的多层感知器(MLP)、径向基函数(RBF)以及分类树模型,以预测学生对于在线开放课程的满意度。Ⅰ型和Ⅱ型误差以及输入变量重要性用于模型比较和分类准确性判断。经过实验,模型的总体分类准确率高于50%。尽管t检验表明模型之间的差异在统计上并不显著,但是MLP模型为低课程满意度的学习者提供了最高的平均分类精度和最低的错误分类偏好。在所有模型中,在线论坛中学习者互动参与度被认为是预测学习者在线课程满意度高的重要指标。In the development of online courses in the past decade,the satisfaction of online learners has always been an important theoretical hotspot.This research relies on the virtual learning environment(VLE)of a large⁃scale online teaching and learning platform.On the basis of nine online open courses offered by a school in the online learning space,the research develops and trains multilayer perceptron(MLP),radial basis function(RBF)and classification tree models with different activation functions in combination with the online learning behavior analysis data in the learning management system(LMS),so as to predict students′satisfaction by online open courses.TypeⅠand typeⅡerrors,as well as the importance of input variables,are used for model comparison and classification accuracy judgment.After experiments,the overall classification accuracy rate of the model is higher than 50%.Although the t⁃test indicates that the differences among models are not statistically significant,the MLP model provides the highest average classification accuracy and the lowest misclassification preference for learners with low course satisfaction.Among all the models,the interactive participation of learners in online forums is considered as an important indicator in predicting satisfaction of learners in online courses.

关 键 词：大规模在线开放课程(MOOC) 网络学习空间 虚拟学习环境 在线开放课程 满意度评价 神经网络 多层感知器 径向基函数 分类树 

分 类 号：TN911.1-34[电子电信—通信与信息系统]