基于困惑表情识别的学习者听课状态诊断及推荐

Diagnosis and Recommendation of Learners’Listening State in Class for Course Selection Based on Confused Expression Recognition

在线学习中学习者对知识点把握困难,进而导致后续知识难以理解,为了解决这个问题,提出了一种基于困惑表情识别的学习者听课状态诊断及推荐方法。该方法首先对学习者全程听课困惑表情进行识别,获得学习者对每节课学习的困惑度。之后,采用基于内容的推荐方法,经过双层Top-N推荐,对学习者困惑的课堂进行针对性的精准推荐。实验结果表明提出的推荐方法可整体把握学习者的困惑知识点并进行精准推荐,其时间复杂度远远低于一般的Top-N推荐的时间复杂度。

In online learning,learners have difficulty in grasping knowledge points,which leads to the difficulty in understandingsubsequent knowledge.In order to solve this problem,a method of diagnosis of learners’listening state in class and recommendationfor course selection based on confused expression recognition is proposed.This method firstly recognizes the puzzled expression oflearners during the whole class,and obtains the degree of confusion of learners for each lesson.After that,the content-basedrecommendation method is adopted to make targeted and accurate recommendations for the confused classes of learners throughdouble-layer Top-N recommendation.The experimental results show that the proposed recommendation method can grasp learners’confusion knowledge points as a whole and make accurate recommendation,and it’s time complexity is far lower than that of thegeneral Top-N recommendation method.