Research And Implementation Of Learner Personalized Recommendation System Based On Knowledge Graph

In recent years,due to the rapid development of online education,the digital resources of various online learning platforms have exploded.How to accurately select relevant and useful information from the massive digital resources is one of the important ways to activate the potential of data and improve the user stickiness of online learning platforms.A recommendation system is undoubtedly an important method to help learners accurately select relevant and useful information.Therefore,based on the analysis of mainstream learning platforms on the current market,this article addresses the issues of rigid learning modes,poor scalability,and lack of resources in personalized recommendation of online learning platforms.Combining the advantages of knowledge graph in feature representation in the recommendation field,using current mainstream development technologies,a personalized recommendation system for learners based on knowledge graph is designed and implemented.This system can provide personalized recommendations for learners,provide personalized learning resources for thousands of people and faces,and also promote and apply them in different courses.The main tasks completed are as follows:(1)To build the knowledge graph of the curriculum,this paper uses the manual construction method to build the knowledge graph.Taking the discrete mathematics curriculum as an example,it realizes the visualization of the knowledge graph,and applies the constructed discrete mathematics knowledge graph to the system.The discrete mathematics curriculum knowledge graph contains 428 entity classes and 382 relationships between entities.This method has scalability and can be used to construct knowledge graph for other courses.(2)A new model for calculating the difficulty of exercises and the mastery of knowledge points was proposed,and a personalized recommendation algorithm for learners based on a knowledge graph was constructed by combining it with a cognitive diagnostic model.This algorithm recommends the most suitable exercises and learning paths for learners to improve their learning efficiency and performance.Specifically,the calculation of exercise difficulty is based on the diversity of exercise difficulty using a combination of classical test theory and question type classification;The calculation of knowledge point mastery takes into account both the learners’ personality characteristics and learning progress.The experimental results demonstrate the accuracy and effectiveness of the recommendation results,demonstrating the effectiveness and rationality of the proposed method.(3)System implementation.This article analyzes the problem of insufficient personalized recommendation in current online learning platforms,and combines the needs of learners.Based on the course knowledge graph and personalized recommendation methods as the core idea,a personalized recommendation system based on the knowledge graph is designed and implemented.The system is developed using mainstream technology stacks,and comprehensive testing is conducted to ensure the feasibility of the system.