Research On Learning Resource Recommendation Based On Topic Mining And Sentiment Analysis

With the rapid development of massive online open courses(MOOC),more and more people are beginning to accept and use online learning methods to acquire knowledge,and online learning platforms have also rapidly grown.On various learning platforms,massive amounts of learning resources and interactive data are continuously produced with the increase in the number of learners.In the face of these data,how learners choose appropriate learning resources has become one of the factors that trouble their participation in online course learning.It takes a lot of time for learners to find learning resources that match their learning style and learning needs,which increases the learner's learning burden and reduces learning efficiency.The course discussion text is the learner's direct expression of the course knowledge content,course setting and teaching implementation,and it is the embodiment of the learner's true learning state and learning needs.By mining and analyzing course discussion texts and recommending resources for learners,it is helpful to improve learning efficiency,quickly and accurately match learners' learning needs,and better achieve personalized education.The research process of this paper mainly includes: firstly,through the web crawler,obtain the course discussion text data generated by the learners participating in the online interaction process.Perform preprocessing such as deduplication,noise reduction,and stop words and word segmentation on the obtained data to form the corpus required for the experiment in this paper;Secondly,collect and organize an sentiment dictionary based on course discussion;Thirdly,use LDA Topic models and dictionary-based sentiment analysis methods enable topic mining and sentiment analysis of learner discussion texts;Fourth,recommend appropriate learning resources for learners based on the obtained topic-sentiment analysis results.The main research work and innovations of this article include:(1)Constructed an online course-based sentiment dictionary,this dictionary including general sentiment dictionary,network sentiment dictionary,domain sentiment dictionary and degree adverbs and negative words sentiment dictionary.(2)Unlike sentiment analysis of all texts,this article mines different topics in the text and conducts sentiment analysis on each topic.The results are more detailed and targeted.(3)Faced with the "resource overload" phenomenon of online learners,this paper proposes a topic and sentiment-based learning resource recommendation method,mining the sentiment orientation of learners to pay attention to topics from the text level,andrecommending test resources and video resources for learners participating in the course of "Python Language Programming".The research results of this paper show that:(1)Compared with the selected general sentiment dictionary,the sentiment dictionary constructed by this course can realize sentiment analysis of course text data more accurately.(2)The recommendation method proposed in this paper can implement resource recommendation based on the needs of individual learners,improve resource utilization efficiency and learner learning efficiency.In summary,on the basis of the existing research,this article through topic mining and sentiment analysis on the text data of learners participating in the course discussion in the online learning platform to obtain learners' attention topics and sentiment orientation,according to the corresponding relationship between the topics and knowledge points,calculate the mastery of knowledge points,and then recommend the resources to learners,in order to provide new ideas and methods for personalized education of online learning platforms.