【作者】 郭成；

【导师】 王艳清； 刘强；

【作者基本信息】 中南财经政法大学 ， 应用统计（专业学位）， 2022， 硕士

【摘要】 2020年心冠肺炎疫情爆发之时,以在线直播互动式教学为特点的网课迅速普及。网课形式下会出现大量的交互记录,这些信息是在校上课过程中所无法存储的,这些数据也会带来一定的价值。提升教学质量是教育的一个重要目标,但是如何提升确是一个难题。老师和学生之间的交互会带来更好的教学质量,因此本文选择中国慕课平台上数学、计算机、外语三类课程的数据,重点挖掘学生对当前网课的一些负面评价的核心观点,并对不同类别课程的负面评价做了多维度对比,探索学生在网课环境下对所学课程的意见以及对不同类别课程的负面评价差异,这将对网课平台、网课老师、在校老师,甚至教育研究者都会有一些借鉴意义。本文阐述了对课程评论情感标注和细粒度评论挖掘相关理论,然后对中国慕课平台数学、计算机、外语三类课程评论进行数据挖掘与实证研究,主要分为六个部分进行课程评论的挖掘。第一部分对网课评论研究的背景和意义以及当前现状进行了阐述,然后对国外和国内关于情感分析和更细粒度的信息抽取进行了文献综述。第二部分对文本处理、情感分析、评价对象-情感词对抽取的理论进行了介绍。第三部分对慕课评论数据如何获取、清洗、转换、可视化分析进行探讨。第四部分是对获取的数据进行极性标注,先基于已有的DUTIR词汇库构建慕课课程评论的情感词典并进行初步极性标记,然后利用情感词典的标记结果使用机器学习的方法对未标记的评论进行极性标记。第五部分抽取负面评论中的评价对象-情感词对并且对评价对象和情感词做了层次聚类。然后通过加权频数、频率、点赞数加权频数、影响度指标对评价对象-情感词对进行横向和纵向的对比分析。第六部分是本文分析得出的结论以及为老师、平台提供的一些意见。本文重点关注数学、计算机、外语课程中负面评论中细粒度的的评价对象和情感词。研究认为,数学类和计算机类课程的老师应当首先解决念PPT和课件的问题以提升课程评分,这两类的老师也应该注意自己的口音,建议使用标准的普通话减少学生在老师口音上的负面情绪,学校在筛选老师时也不能忽略普通话的要求。计算机和外语类课程需要增加课程内容的趣味性,做到灵活多变而不失条理,尽量不要学生感到课程和内容枯燥无味,同时也要更加关注学生的基础情况,做到课程内容和作业安排因人而异。更多还原

【Abstract】 When the epidemic of coronary heart disease broke out in 2020,online classes characterized by live online interactive teaching were rapidly popularized.In the form of online class,there will be a large number of interactive records.These information can not be stored in the process of school,and these data will also bring some value.Improving teaching quality is an eternal topic of education,but how to improve it is a difficult problem.I think the interaction between teachers and students will bring better teaching quality.Therefore,this thesis selects the data of mathematics,computer and foreign language courses on the Chinese MOOC class platform,focuses on mining the core views of students on some negative evaluations of current online courses,and makes a multi-dimensional comparison of the negative evaluations of different types of courses,It is believed that it will be helpful to explore the students’ opinions on the courses they have learned and the differences in the negative evaluation of different types of courses in the online class environment.This thesis expounds the relevant theories of emotional tagging and fine-grained comment mining of curriculum comments,and then makes data mining and empirical research on three kinds of curriculum comments: mathematics,computer and foreign language.It is mainly divided into six parts to mine curriculum comments.The first part expounds the background,significance and current situation of online class comment research,and then summarizes the literature on emotion analysis and finer grained information extraction at home and abroad.The second part introduces the theory of text processing,emotion analysis and evaluation object emotion word pair extraction.The third part discusses how to obtain,clean and visually analyze MOOC comment data.The fourth part is to label the polarity of the obtained data.Firstly,the emotional Dictionary of Moke course comments is constructed based on the existing dutir vocabulary and marked with preliminary polarity,and then the unmarked comments are marked with polarity by using the marking results of the emotional dictionary.The fifth part extracts the evaluation object emotional word pairs in the negative comments,and makes a hierarchical clustering of the evaluation object and emotional words.Then,it makes a horizontal and vertical comparative analysis of the evaluation object emotional word pairs through the weighted frequency,frequency,praise number,weighted frequency and influence index.The sixth part is the conclusion of this thesis and some suggestions for teachers and platforms.This thesis is not limited to the emotion of the subject,and obtains some positive and negative emotion intensity,but focuses on the finer granularity of the evaluation object.The research shows that teachers of mathematics and computer courses should first solve the problem of reading PPT and courseware to improve the course score.Teachers of these two types should also pay attention to their accent and suggest using standard Mandarin to reduce students’ negative emotions on the teacher’s accent,schools can not ignore the requirements of Putonghua when screening teachers.Computer and foreign language courses need to increase the interest of the course content,be flexible and organized,try not to make students feel boring,but also pay more attention to the basic situation of students,so that the course content and homework arrangement vary from person to person.更多还原