Research On Recommendation Of Personalized Exercises Based On Knowledge Graph

With the rapid development of Web2.0 and Internet education,flexible and freely customizable online exercises have begun to spread among learners.At present,most online learning platforms lack the consideration of information overload,which leads to problems such as less targeted and inefficient learners.Faced with a large number of problem data in the network and in the examination system,it is a hot issue in the field of personalized education to consider how to make up for the knowledge loopholes and to select exercises to promote the effective learning of learners.The problem recommendation technique is an important way to solve the problem of problem information overload,and gradually infiltrates and changes the learning style of many learners.Personalized problem recommendation system as an important branch in the field of education,in recent years,with the development of artificial intelligence and big data has received great attention from educators.Based on the lack of consideration of the relationship between the knowledge points of most courses,this paper establishes the curriculum knowledge map as the knowledge model based on the relationship between the knowledge points and knowledge points,and studies the problem recommendation algorithm based on knowledge map.On this basis,this paper simulation experiment of the proposed algorithm proposed in this paper was carried out using the exercises of the junior high school mathematics course.The specific research content of this paper mainly includes the following four aspects:1.Based on the relationship between knowledge points and knowledge points,the construction and storage methods of knowledge models are proposed.Firstly,the extraction method of the concept is studied,and the knowledge points of the course are extracted using the related techniques of natural language processing.Then the relationship between knowledge points is analyzed and the relationship between knowledge points is designed.The curriculum knowledge map is constructed based on the relationship between knowledge points and knowledge points.Finally,the storage method of knowledge model is studied.Finally,the storage method of knowledge model is studied.2.Based on web crawler technology and database technology,the crawling and preprocessing techniques of network exercises are studied.By studying the characteristics of different problem types,the multiple choice questions were selected as the problem types suitable for the recommendation system.2.Then,build a storage model of the problem,use the data storage technology to store the problem in the database as a problem library.3.The problem recommendation algorithm based on knowledge map is studied.This algorithm uses knowledge tracking technology as a method for judging learners' mastery of knowledge points.It not only informs learners that the knowledge points that have not yet been mastered are caused by the lack of mastery of those knowledge points,but also guide learners to master a certain knowledge.Based on the point,the next step should be to choose which knowledge point to learn,and then recommend the corresponding problem to the learner according to the backward recommendation strategy,so as to deepen the learner's understanding of the knowledge point.4.Taking the junior high school mathematics problem as an example,the experimental simulation method based on the knowledge map is verified by experiments.The experimental results show that the problem recommendation strategy based on knowledge point relationship is an important auxiliary learning method,which not only can make up for the knowledge loopholes of learners with different learning abilities,but also improve the learner's independent learning efficiency.