Research On Personalized Learning Path Recommendation Method Base On Knowledge Graph

In the context of the rapid development of education informatization,the total amount of digital education resources has increased exponentially.Offline education is no longer the only way for learners acquiring knowledge.Online education realizes users’ self-learning and expands the way of knowledge acquisition.But the following question is how to realize the personalized acquisition of resources user demanded and systematic learning of courses in this environment the scale of digital education resources sharply increase.Although the current search engines use inverted index to improve the hit ratio of knowledge entity search and the efficiency of learners in acquiring education resources,learners still face difficulties in correct orientation of learning and systematic learning of courses after completing the current stage of learning.Therefore,personalized learning path recommendation in line with the logical structure of subject knowledge has become a research focus of online education,with strong epochal and development value.In order to improve the systematization of knowledge acquisition,this paper studies the personalized learning path recommendation based on knowledge graph,and proposes a method to realize the orientation of learning direction and personalized recommendation after accurately understanding user preferences and learning styles by mapping knowledge entities based on the relationship between knowledge entities.Personalized learning path recommendation system is designed and implemented based on relevant research.The main contents of this paper are as follows:(1)Construct user portraits for learners’ personalized learning paths,and obtain learning styles by proposing an adaptive learning style generation method analyzes user behaviors on the basis of acquiring explicit learning styles,adjusts learning styles according to rules and improves the authenticity and timeliness of portraits;Using BERT+BI-LSTM+CRF model to improve the accuracy of interest keywords extraction,accessing user interest entities better;The embedding algorithm of vector representation learning is improved to optimize the training time and speed.(2)A personalized learning path recommendation algorithm based on knowledge graph is proposed by improving Page Rank algorithm.The weight of each relationship coefficient is adjusted according to learning style,and the importance coefficient of nodes corresponding to different learning styles is obtained.Personalized learning path is generated after sorting.The feasibility of the algorithm is proved by analyzing the results of path generation on three data sets,and the availability of the algorithm is proved by comparing the generated path and the offset between catalogue and real path through simulation experiments.(3)A personalized learning path recommendation system is built,which implements modules such as user learning style generation,interest entity extraction,and personalized learning path recommendation.Provids a personalized path recommendation function based on knowledge graphs for digital education platforms,displaying the path in a visual way.This paper studies model construction,algorithm design and system implementation,which provides a solution for personalized learning path recommendation based on knowledge graph and promotes the development of digital education.