Research On Personalized Recommendation Algorithm For Student Learning

Most of the existing learning platforms directly apply the traditional web recommendation algorithm to students’ course learning,and find the course resources with similar content according to the students’ historical learning records to make recommendations.However,such algorithm do not deeply explore the students’ learning characteristics and the social relationship with students.They also face the difficulties of low recommendation accuracy and poor personalized learning effect.For the knowledge points in the same course resources,the existing recommendation algorithms have less research on knowledge points and lack of analysis of students’ learning data.It can not evaluate students’ mastery of knowledge points reasonably,and cannot effectively recommend the required knowledge points for students.In view of the above problems,this paper takes improving students’ personalized learning as the goal,and analyzes the difficulties faced by students in choosing course resources seriously and learning specific course knowledge points.This paper introduces the trust relationship between students,proposes a course resource recommendation algorithm,constructs the course knowledge graph,improves the word vector language model,and proposes a course knowledge point recommendation algorithm model.The main contents and innovations of this paper are as follows:（1）Aiming at the recommendation of course resources,this paper proposes a course resources recommendation algorithm integrating students’ implicit trust.The algorithm defines implicit trust between students as the supplement of explicit trust,and adds trust bias to the prediction formula of students’ course score,so that the trust between students can correct the deviation caused by the recommendation algorithm.The improved similarity calculation formula enables the algorithm to consider both the deviation of students’ evaluation and the common evaluation of course resources.Experimental results showed that the algorithm has high recommendation accuracy even under the condition of sparse data and cold start.（2）Aiming at the problems of the number and types of knowledge points,and the complex relationship between them,this paper proposes a method of constructing and reasoning C++ course knowledge graph based on Neo4 j.This method can effectively extract the course knowledge point entities and the relationship between them,and store them in the graph database Neo4 j in the form of triples.The quality evaluation of the constructed course knowledge graph proves the effectiveness of the construction method.The constructed course knowledge graph strengthens the learning order among course knowledge points,and lays a foundation for the research of course knowledge point recommendation algorithm.（3）Aiming at the recommendation of course knowledge points,this paper proposes a knowledge point recommendation algorithm model based on course knowledge graph.According to the recommendation requirements of knowledge points,the model improves the network structure and feature information fusion strategy of the word vector language model,and solves the problem of feature information extraction is not rich in the course corpus data.According to the actual learning situation of students,a variety of knowledge recommendation strategies are defined,which can effectively recommend personalized knowledge points for students.Finally,compared with other mainstream algorithm models,the effectiveness of the model is proved.