Research On Discourse Style And Content Outline Of Online Courses

With the deep integration of information technology and education and teaching,the barriers of space and time have been broken,and the rapid development of online education has been promoted.However,the massive learning resources make it difficult for learners to choose.How to obtain the learning materials that learners want from the massive data is an urgent problem to be solved.To this end,this study uses 134 online courses of "Speech Comprehension" from ten teachers as research data to analyze the characteristics of online course learning materials and assist learners to achieve personalized learning.First,based on online course teacher discourse style identification and validity verification.Combined with the current natural language processing technology,it mainly includes word segmentation,sentence segmentation,part-of-speech tagging and other methods for data preprocessing to build a teacher’s utterance corpus.By sorting out the relevant literature on quantitative stylistics and combining the relevant characteristics of online courses,an indicator system for teacher discourse style is constructed.This system divides teacher discourse style into five first-level indicators: rhythm,richness,colloquialism,humor and descriptiveness.Using the K-Means clustering method to cluster the 8 teachers’ discourse style characteristics indicators,analyze the discourse style characteristics of various teachers in the clustering results,and classify the discourse styles of the ten teachers in the "Speech Understanding" course as fast and emphatic,rich humor type and natural ordinary type.From the verification of the accuracy of the recognition results,it was found that teachers such as T01 and T05 had relatively accurate recognition results,while T06 and T07 had poor recognition results.Secondly,based on the content outline extraction and validity verification of online courses.Use the gensim package in Python to vectorize the text,use the topic-consistency curve to judge the number of course topics that each teacher should extract,take empirical values for the two hyperparameters,and use the LDA topic model to extract the course topics and topics for each teacher Words,analyze the topic extraction results,and find the differences between the teachers in the course arrangement,content structure,etc.Use natural language processing technology to divide the text into sentences,calculate the similarity matrix between sentences,and combine the Text Rank algorithm to extract the top-ranked sentences as the text summary of this course.Validation was conducted from the two aspects of course subject,accuracy and redundancy of course text summaries,and it was found that T01,T03,etc.were teachers with high accuracy of course subjects,and T08,etc.were mostly teachers’ course text summaries The extraction results are more accurate and less redundant.Finally,based on the characteristics of the teacher’s discourse style and the outline of the course content based on the online course,a course briefing is formed,which briefly summarizes the course characteristics of each lesson,improves the efficiency of learners in choosing learning materials,and promotes personalized learning,digital learning and lifelong education develop.