Research On Personalized Recommendation Method Of Learning Resources Based On Behavior Sequence

With the rapid development of the Internet and information technology,people’s demand for education is increasing,prompting the diversification of modern education methods.Currently,many online education fields are providing learning services to users by integrating high-quality courses and other online course resources.However,in the face of massive learning resources,the problem of information overload has become more and more serious,and it is difficult for users to find the right learning resources for themselves in a targeted manner,and they even have problems such as cognitive load.Recommendation systems are widely and successfully used in e-commerce to accurately recommend products of interest to users.Recommender systems can solve the information overload problem well,and the recommendation algorithm commonly used in the e-commerce field is introduced to the online education field,which can provide personalized learning resource recommendation service according to users’ interests.At present,the application of recommendation of online learning resources in the education field is still in its initial stage,and the use of traditional recommendation methods mainly explores the static correlation between users and learning resources,but ignores the dynamic changes of users’ interests.On the other hand,considering only the historical interests of users cannot accurately express the diversity of users’ interests,which leads to poor recommendation results.To address the above problems,the main work of this paper is as follows.(1)For the shortcomings of traditional recommendation methods to mine the static correlation between users and learning resources,a personalized recommendation model based on Attention mechanism and users’ Long and Short term Learning Interest behaviors(ALSLI)is proposed.The user’s behavior sequence is divided into long-term and short-term sequences,and the attention mechanism is used to extract the long-term and shortterm interests of the user,while the user’s interest expression is obtained by fusing the user’s characteristic attributes.The experimental results show that the sequence model incorporating the attention mechanism and the long-and short-term interests can more accurately simulate the user’s interest changes,which improves the accuracy of the recommendation results.(2)For the problem that existing recommendation models using a single interest vector cannot accurately express the diversity of users’ interests and are not sensitive enough to extract learners’ multifaceted interests,which leads to poor results in recommending learning resources,this paper proposes a learning resource recommendation method based on Capsule networks and Learning Styles for learners’ Multifaceted Interests(CLSMI),using a dynamic routing algorithm in capsule networks to extract users’ multifaceted interests,introducing The Felder-Sliverman learning style quantification table is introduced to model the learning style preferences of users.By comparing the analysis with related multi-interest extraction models,the method used in this paper obtains better results and also improves the user experience.(3)In this paper,the recommendation algorithm based on user behavior is applied to the field of online education to provide some new ideas for the development of the field.On the basis of the above two recommendation models,an online learning website is designed and developed to put into application,which contains services such as video on demand,course purchase,personalized learning resources recommendation to meet the needs of users’ fragmented learning in real life and make the research have practical application value.