Personalized Learning Resources Recommendation Model And Application For Students

With the rapid development of information technology and data science,as well as the concept of "Internet +" education innovation,personalized E-learning has received extensive attention in school education and home education.With the development of education informatization,the number of E-learning users is increasing rapidly,and the number of learning resources is also appearing at an explosive speed,which makes learners face the dilemma of "information overload" and "learning confusion" in the process of learning.Faced with massive amounts of learning resources,how to tap and push learning resources suitable for the development of learners in a limited period of time is crucial to achieve "individual teaching" and "individual development."To meet the individual needs of learners,a personalized learning resource recommendation system comes into being.We recommend appropriate learning resources,learning environments and learning partners for learners according to their characteristics.In the personalized learning resource recommendation system,the construction of learner model is the most important.At present,most learner models have the following problems: the dimension is not scientific enough,the feature attribute acquisition method is single and the representation method is not computable.These problems can cause the learners' learning ability not to match the difficulty of the recommended learning resources,and even lead to the learners' cognitive overload or loss in the process of learning.The paper aims to provide a strong support for personalized learning resource recommendation by constructing learner model and learning resource model,and to effectively alleviate the cold start and sparsity problems in the process of recommendation.The main contents of this paper are as follows:(1)Construction of learner preference model.Through the analysis of learner's behavioral data in the process of learning,the learner model is analyzed by extracting the three cognitive features of learner's cognitive ability,knowledge level,and learning resource preference.Among them,a detailed study is mainly focused on learning preference information,and Ontology is used to construct a learner preference model to better realize the semantic relationship between the knowledge so as to discover students' interest in learning.(2)construction of learning resource model,and refinement of personalized learning content.Based on Ontology,a learning resource metadata description model and a middle school knowledge point model are constructed and described in OWL language.(3)Reaearch on personalized learning resource recommendation model.By analyzing the problems of poor portability and high complexity of learning resource types in learning resource recommendation methods,this paper chooses a collaborative filtering recommendation algorithm.Based on the dual cluster analysis of learners and learning resources,the collaborative filtering recommendation algorithm is improved and personalized learning resources are recommended.The experimental results show that the double clustering collaborative filtering recommendation algorithm is more effective.