Research On Personalized Recommendation Of Learning Resources Based On Learning Behavior

With the continuous integration and further development of Internet technology in the education field,online learning platforms have brought new learning experiences and learning methods to the general public.Online learning has many advantages such as abundant resources and easy access.However,due to the problems of the Internet itself,such as "Knowledge Lost" and "data flood" are becoming more and more prominent,how to recommend resources according to the characteristics of users,so as to effectively improve the personalized learning experience,has been a subject of extensive research by researchers.While learners are learning on the current common learning platform,the course pages and resource recommendations are the same,which requires learners to choose their own courses and access paths,sometimes some questions cannot be answered in time,such learning platform lacks the initiative to push personalized learning support services to learners,and cannot meet learners ' personality and emotional needs.Therefore,it is very necessary to study the personalized recommendation of learning resources in the online learning platform,and it is also very useful for practical reference.In view of the above problems,this paper designs a recommendation method based on the collaborative filtering of users' recommendation idea,which is based on the solution of commodity recommendation in the field of business,combined with learner data.The main research contents of this paper are as follows:Firstly,summarized the present situation of learning resources recommendation?the research status of peer mutual aid learning and the current situation of peer communication in learning platform by literature research.The research shows that online learning platform pays more and more attention to personalized experience and emotional experience,but the number of literatures recommended by learning peers is far less than the number of curriculum recommendations.Secondly,based on the learner dataset of Canvas Network platform,the learner data is divided into basic population characteristic data and learning behavior data,and then the peer recommendation model is established.Then,the correlation analysis of learner behavior data is carried out to determine the behavior variables in the clustering,and the one-way variance analysis of the basic population characteristics data is carried out to determine the classification variables in the cluster,and the learner cluster analysis is carried out by synthesizing the two kinds of variables.According to the characteristics of the data,the appropriate recommendation algorithm is selected,and the weights of the behavior variables are assigned by the information entropy.The main function of analyzing data is to extract behavioral features and provide an objective data basis for later decision making.Finally,based on the collaborative filtering recommendation idea,according to the clustering results of peer recommendation to learners,and using Euclidean distance algorithm to calculate the distance between the same learners,to the learners with less learning behavior recommended behavior more learners.The results are sorted in ascending order and presented by TOP-N.