Knowledge Association Analysis And Knowledge Recommendation Based On Course Data

Digital teaching has been emerging in the field of higher education at home and abroad with the development of information society.Especially in early 2020,the outbreak of new crown pneumonia was restricted.Online education was restricted and online education was further popularized.This provided more abundant data resources for Educational Data Mining(EDM)research.EDM uses machine learning technology to discover knowledge from education data,aiming to improve the level of students' learning,teachers' teaching and education administrators' management.At present,most of the curriculum-themed research in EDM focuses on using traditional machine learning algorithms to mine the relevance of courses and predict students' academic performance.However,due to the long time span of the course,it is not conducive to the education administrators to adjust the teaching plan in time.At the same time,the granularity of course connection is relatively large in the field of students' learning,and it is unable to provide students with specific learning help after finding problems.Therefore,this thesis takes the "program design" course as an example,the knowledge points in the course as the research object,clarifies the key points of the course by mining the association of knowledge points,clarifies the difficulties of the course by classifying and analyzing the students and knowledge points,and recommends the knowledge for the students through collaborative filtering.Finally,the goal of assisting teachers in teaching and guiding students in learning can be achieved.To sum up,the main work of this thesis is as follows.1.This thesis analyzes the domain model and data model of the program design examination system of Software Engineering College of C University from abstract to concrete,finds out the required data from the scene through the instantiation of classes,and preprocesses the data,so as to provide the basis for the follow-up work.2.Based on the association pattern mining method and support confidence framework,the association rules between test questions are mined,and the knowledge points associated with test questions are analyzed.The knowledge association network diagram is obtained,and the key knowledge of the course is defined.3.Based on the improved clustering algorithm,the students are classified according to their grades and knowledge points according to their difficulty.The mastery degree of different kinds of students on knowledge points with different difficulty is analyzed,and the difficulties of different students in learning this course are clarified.Finally,some suggestions are put forward for students of different levels to study the course.4.Based on the needs of students' course knowledge recommendation,the collaborative filtering algorithm combined with association rule filling is improved.The experiment shows that the effect of the improved algorithm in this application scenario is better than the traditional algorithm and the original algorithm.The improved algorithm is applied to the actual scene to make clear the students' knowledge level,and recommend the poor knowledge points to the students.