Personalized Learning Resource Recommendation With Cognitive Diagnosis

With the continuous advancement of educational informatization,the ways of learners to obtain learning resources are gradually diversified,from the traditional active retrieval way to the system automatic recommendation way.The intelligent behavior of learning content recommendation significantly improves the efficiency of users accessing personalized resources;With the rapid development of network information technology and big data technology,personalized online learning of learners has been widely concerned by the education industry.The number of all kinds of learning resources on the Internet is increasing explosively.Learners are also confronted with the dilemma of "information overload" and "learning trek" in the process of choosing online resources for learning.In the face of massive online learning resources,how to accurately and effectively excavate learning resources suitable for learners is the challenge facing the future education towards the field of information technology.(1)In view of the problems of blindly following learning and low learning efficiency caused by information and knowledge overload caused by massive learning resources at present,this paper firstly introduces the theories related to personalized learning and cognitive diagnosis,makes a comparative analysis of the current research status,and defines the research direction and objectives of this paper.It provides a theoretical basis for the following consideration of personalized learning resource recommendation from the aspects of learners’ learning defects.(2)This paper proposes a similarity calculation method based on learners’ mastery of knowledge points.Based on Pearson’s similarity calculation formula,this method introduces three similarity factors and improves them.Combined with the prediction method of K-nearest neighbor algorithm,the similarity of different learners’ learning situations is calculated,thus improving the accuracy of similarity calculation results.Based on this model,learners can recommend personalized learning schemes to learners according to their own learning conditions.Learners identify important knowledge points to be learned in the knowledge structure diagram according to the learning path set in the scheme and the recommendation degree of each knowledge point in the knowledge structure diagram,so as to recommend more appropriate learning resources to learners.By comparison with the experimental verification,the proposed scheme model,compared to other models,has obvious improvement to the promotion of learners’ learning.(3)Based on the correlation between course knowledge points,this study constructed the structure chart of course knowledge points,and combined with the cognitive diagnosis method,calculated and reasoned the exercise data of learners,obtained the learners’ mastery of knowledge points,and generated the sequence of defective knowledge points,which was an important basis for the recommendation results.Secondly,the degree centrality of knowledge points is introduced into the knowledge point structure diagram to calculate the importance of each knowledge point in the course knowledge point structure diagram,as one of the basis for recommending learning resources to learners.In contrast to previous studies on learning resource recommendation,learners’ mastery of exercise knowledge points was not considered.In this study,learners’ mastery of exercise knowledge points was calculated from the perspective of exercise-answering,so as to diagnose their knowledge level.It solves the problem that the diagnosis of learners’ knowledge level by the existing cognitive diagnosis method only focuses on whether the learners answer the exercises correctly or not,and also provides theoretical and data support for the recommendation of personalized learning resources for learners in the future.Finally,through experimental comparison and verification,compared with other recommendation models,the recommendation model proposed in this paper has significantly improved learners’ learning performance and learning effect.