Research On The Key Technology Of Adaptive Learning System Based On Knowledge Graph

Personalized learning has always been the goal of educational reform and development in all countries of the world.From the principle of "teaching students according to their aptitude" proposed by Confucius to the idea of "learner-centered" advocated by Dewey to the current idea of "wisdom education",personalized learning is regarded as the essential demand and value orientation of promoting students’ development.In the traditional industrial era,the education system based on class teaching and standardized teaching materials can no longer meet the personalized needs of the current society for education.The adaptive learning system supported by information technology has become the "experimental field" and breakthrough to achieve the goal of personalized learning.Adaptive learning system is usually built on domain knowledge model,learner knowledge model,adaptive engine,its key technology includes domain knowledge modeling,learner knowledge modeling,personalized learning path recommendation.Although existing research has made certain achievements,there are still some shortcomings,mainly displays in: traditional domain knowledge models based on concept map,knowledge map and cognitive map can not express a wide range of knowledge types and rich semantic relationships,while the existing educational knowledge graph has problems of unclear model and low construction efficiency,the accuracy of knowledge tracking and personalized learning path recommendation is not high;in addition,there is no a comprehensive system of evaluation index system.In view of the above problems,this study takes knowledge graph as the breakthrough point,and makes a deep study of several key technologies in the construction of adaptive learning system,and preliminarily obtains the following research results:(1)A conceptual model of educational knowledge graph oriented to adaptive learning system and its intelligent construction method are proposed.Aiming at the problem of unclear model and low construction efficiency of educational knowledge graph,this study first constructs a conceptual model of educational knowledge graph for adaptive learning system based on cognitive map,knowledge space theory and cognitive diagnostic test theory.On this basis,the method for knowledge element extraction combining rules and machine learning is proposed,and the method for precursor-successor relationship mining based on association rule algorithm is proposed.This research lays a foundation for the follow-up study of knowledge tracking and personalized learning path recommendation.(2)Deep knowledge tracing model supported by educational knowledge graph is proposed.Aiming at the problem that the accuracy of learner knowledge tracking is not high,this study proposes a deep knowledge tracking model—DKT-KG model,which integrates domain characteristics with the support of educational knowledge graph.The model filters important evaluation behavior characteristics through decision tree,and combines the knowledge dependence relationship represented by educational knowledge graph.Experiments on ASSISTment 2009-2010 data show that the model has high validity and accuracy.The knowledge level evaluated by the model provides an important basis for the recommendation of personalized learning paths.(3)A method for personalized learning path recommendation supported by educational knowledge graph is proposed.Aiming at the problem that the accuracy of personalized learning path recommendation is not high,this study first proposes the idea of personalized learning path recommendation supported by educational knowledge graph,and then proposes a learning path generation method based on improved ant colony algorithm on the basis of describing the characteristics of learners and learning objects.The experimental results show that the proposed method has significant improvement in recommendation quality and algorithm performance.This study provides a powerful guarantee for achieving the precise personalized learning goal.(4)The prototype of adaptive learning system based on knowledge graph is designed and implemented.In order to further verify the feasibility of the above key technologies,according to the theory and method of software engineering,the supporting system was designed and implemented.Firstly,the system is analyzed from the perspective of functional and non-functional requirements,and then the overall design was carried out from the overall logical architecture,system functional structure and data model.Then the interface and implementation class are designed for five functional modules: user management,resource management,evaluation management,knowledge graph and learning path recommendation.Finally,the main functional modules are implemented.The system provides systematic support for the implementation of personalized learning.(5)A comprehensive evaluation index system of adaptive learning system based on knowledge graph is constructed.In order to effectively evaluate the quality of the developed system,this study takes the successful model of information system and the technology acceptance model as the theoretical basis,uses Delphi method and analytic hierarchy process to construct the corresponding evaluation index system,and then evaluates the system through the fuzzy comprehensive evaluation method.The results provide a reference for the improvement and optimization of adaptive learning system.The contributions of this research are mainly reflected in the following two aspects:(1)On the technical level,firstly,the conceptual model of educational knowledge graph Oriented to adaptive learning system and its intelligent construction method are proposed;secondly,the deep knowledge tracking model and personalized learning path recommendation method supported by educational knowledge graph are proposed,which provide technology for personalized learning based on adaptive learning system.(2)From the perspective of the integration of natural science and social science,this study adopts the combination of qualitative and quantitative analysis methods,and the idea of software engineering to study the key technology of building adaptive learning system based on knowledge graph,which provides some reference for the perspective selection of the future personalized learning research.