| Along with the popularization of intelligent mobile terminal,the development of Cloud technology and the arrival of 5G communication technology and the explosive growth of shared learning resources,online learning has become a trend.In order to meet the learning characteristics,efficiency and pertinence of different learners' online learning,personalized recommendation function is particularly important for the development of online learning system.The structured organization and representation of curriculum knowledge is a powerful guarantee for the realization of personalized learning recommendation system.At present,there are more researches on curriculum knowledge graph of basic education in primary and secondary schools,while there are fewer researches on the construction and visualization of the curriculum knowledge graph in university,and the construction of the curriculum knowledge graph has not formed a unified standard."Data Structure" mainly introduces and studies the storage and processing methods of data in computer,which aims to develop students' ability to analyze data,organize data and guide students to write efficient and structured programs.In addition,"Data Structure" is a basic course in specialties of computer,electronic information,automation.Therefore,this paper chooses this course as the research object,studies the theory and technology of knowledge graph construction,and designs and implements a personalized exercise recommendation system based on knowledge graph.The main research contents are as follows:(1)Construct curriculum knowledge map and store it.This paper studies the theory and technology of knowledge graph construction,designs the top-level ontology of curriculum,uses the ontology visualization tool Protégé to build the curriculum ontology of data structure,obtains the examples of curriculum ontology and its relationship,and finally constructs the curriculum knowledge map of data structure.The popular graph database neo4 j is used to store the curriculum knowledge map.In the graph database neo4 j,knowledge visualization,the queries of knowledge point,knowledge attribute and knowledge shortest path are realized.(2)Design exercise classification and propose the algorithm of exercise recommendation based on exercise classification.This paper uses naive Bayes and support vector machine algorithm to classify exercise texts and makes experimental comparative analysis,and uses the advantages of support vector machine algorithm to deal with small sample data well.This algorithm is selected to automatically classify exercises according to chapters and mark chapter categories for each exercise.Then,combined with the semantic relationship of knowledge graph,we design exercise recommendation algorithms based on section(ERS)and knowledge point(ERKP),and recommend the exercises that meet their cognitive level and are of moderate difficulty to learners through exercise test,which can achieve personalized recommendation of exercise.(3)Design and implementation of exercise recommendation system.Based on the B/S architecture,the overall structure,functional modules and relational database tables of the system are designed.While the design of the underlying database is mainly the combination of neo4 j and MySQL,and technology of d3.js is used to visualize course knowledge graph,to realize the queries of knowledge point and knowledge path,and to realize the exercise recommendation algorithm based on ERS and ERKP,which can achieve the effect of personalized exercise recommendation.The personalized exercise recommendation system based on knowledge graph is applicable to learners in each learning stage.As long as learners click the relevant knowledge nodes in the graph,they will be push to the relevant exercises and to test.According to the level of test scores,the weak knowledge points of learners are diagnosed and personalized exercise recommendation is be carried out to improve learners' learning efficiency. |
PDF Full Text Request