Research On Personalized Recommendation Of Online Learning Resources Based On Knowledge Graph

With the rapid development of information technology and the Internet,people have gained unprecedented free information space.Everyone is the information collector and information producer,which also makes the information in a state of explosion.Text,pictures,video,audio and other contents gathered under various information platforms form a variety of modes,update and fast big data,and gradually produce "information overload" and "resource maze".How people find the information and resources they really need in the mass information has gradually become a problem.Compared with the search engine,the recommendation system does not require users to actively give search keywords to obtain information,but can actively build interest model according to the user’s historical preference information,and make timely and accurate personalized recommendation for users,so as to better meet the user’s demand for information,so it is more concerned by experts and scholars.However,in the start-up phase,the recommendation system often has the problem of cold start,and the user history is often sparse,which makes the performance of the recommendation system limited.Data is widely available,what is lacking is the ability to extract knowledge from it.Knowledge graph is the most important form of knowledge representation in the era of big data.As a priori knowledge,it can provide semantic features for recommendation algorithm.The introduction of knowledge graph into the construction of recommendation model can effectively alleviate the cold start and data sparsity problems,and help the recommendation system achieve better performance.In addition,with the popularization of mobile terminals and the prevalence of online learning platforms,various flexible education methods can help learners to use fragmented time for different forms of learning according to their own needs,and the advantages of personalized recommendation can also be effectively reflected in the online learning system.Mining the characteristics of learners and learning resources,studying the matching relationship between them,comprehensively using knowledge graph,recommendation algorithm,association analysis,logical reasoning and other methods to provide personalized learning resources recommendation services for learners,forming the application framework of online learning personalized services,which is an important development direction of the ultimate goal of knowledge service.From the perspective of the construction and utilization of educational knowledge graph,this study aims at the knowledge service for learners,and proposes the personalized recommendation method of online learning resources based on knowledge graph combining the characteristics of learners and learning resources.This research divides the personalized recommendation system of online learning resources based on knowledge graph into three dimensions according to the three levels of user needs: learning objectives,learning motivation and learning planning,combined with the three dimensions of learner profile information of interest,behavior and status,and three collaborative filtering recommendation methods based on project,user and model Project collaborative recommendation based on learning interest analysis,user collaborative recommendation based on learning behavior analysis and model collaborative recommendation based on learning state analysis.In each recommendation dimension,appropriate knowledge graph instances and ontology architecture are used to mine and express the characteristics of learners and learning resources.In the four elements of education knowledge graph structure based on "subject curriculum knowledge point learning resources",collaborative filtering recommendation methods of different dimensions are implemented and experimental design and result analysis are carried out to form online learning resources based on knowledge graph source personalized recommendation application framework.This paper is a study of theory and method for practical application.The paper consists of 7 chapters,including 51 charts and 10 tables.The main contents of each chapter are as follows:Chapter 1: introduction.This paper summarizes the research background,significance and research status of the topic,and through the overview of the main content,leads to the research ideas,research methods and technical routes,which plays a role in guiding the whole paper.Chapter 2: theoretical part..First,through a large number of references,combing and comments on relevant research at home and abroad,this paper expounds the construction theory of knowledge graph and related extension knowledge.The second is to systematically sort out and summarize the common personalized recommendation methods,analyze their development status and shortcomings,and provide theoretical basis for the personalized recommendation methods proposed later.Thirdly,it summarizes the related theories and methods of information mining involving interest pattern,behavior pattern and personality pattern.Chapter 3: the design and construction of ontology structure of education knowledge map.Starting from the basic logic of curriculum teaching and the functional requirements of adaptive learning system,an adaptive and extensible ontology model of curriculum knowledge representation is designed based on knowledge map technology,which provides a solid foundation for the following chapters.Chapter 4: project collaborative recommendation method based on learning interest analysis.Based on the introduction of subject knowledge map and the prediction and analysis of learners’ interest through a small amount of feedback,the paper proposes to integrate the information amount of resources into the traditional semantic relevance similarity,so as to realize the beneficial transfer and expansion of user interest preference,improve the project recommendation specialization,and provide diversified learning resources based on the analysis of learners’ interest Source.Chapter 5: user collaborative recommendation method based on learning behavior analysis.Based on the introduction of resource knowledge graph,this paper analyzes the implicit behavior of learners’ behavior log,considers the explicit scores of learning resources comprehensively,complements the sparse score matrix through the comprehensive scores formed by the display scores and implicit scores,and proposes an improved user similarity calculation model to better improve the recommendation effect.Chapter 6: model collaborative recommendation method based on learning state analysis.Based on the introduction of curriculum knowledge graph,the knowledge layer and cognitive layer based on curriculum knowledge points are constructed.Through the judgment of learning style,diagnosis and evaluation of cognitive level of learners in the learning process,the adaptive learning activity sequence is displayed in a visual way,which is the basis of personalized learning path and improves recommendation satisfaction.Chapter 7: summary and prospect.This paper summarizes the main points of the personalized recommendation method based on knowledge graph in the online learning environment,expounds the value,innovation and management enlightenment of this study again,and explains the shortcomings and limitations of this study,as well as the future research work.Learner-centric personalization of online education is already in full swing,and future research and application of knowledge graph technology-based approaches will play an increasingly important role at the platform,enterprise,government and even national levels.This study is a useful exploration in an interdisciplinary field,and it is hoped that it will serve as a point of reference and reference for knowledge services oriented towards online learning.