Collaborative Filtering Based On Item Popularity Weighting Research And Development Of Recommendation System

With the rapid development of information technology on the Internet,a large number of online courses have emerged,providing a wealth of learning resources and flexible choices for many learners.At the same time,the phenomenon of information overload on the Internet platform is becoming more and more serious,which leads learners to spend a lot of time and energy searching and choosing learning resources suitable for themselves,and the results are often unsatisfactory,which greatly weakens the auxiliary role of online learning platform.To solve this problem,collaborative filtering recommendation method can be used to recomnend personalized learning curriculum resources to users.Based on the analysis of popularity recommendation algorithm,a popularity recommendation algorithm with user grouping is proposed to solve the cold start problem of new registered users in collaborative filtering recommendation technology.For the old users,the difference of popularity and popularity has a great influence on the calculation of similarity in collaborative filtering.A collaborative filtering recommendation algorithm based on adaptive popularity threshold is proposed,which combines item popularity weight with adaptive popularity threshold.By setting adaptive popularity threshold and weight of popularity penalty and compensation,it is introduced into the similarity calculation formula.For courses whose popularity is greater than the threshold,the recommendation algorithm is proposed.Punishment is applied to compensate the courses whose popularity is less than the threshold,which reduces the influence of the difference between popularity and popularity on the calculation of similarity of collaborative filtering projects,and improves the coverage of personalized curriculum recommendation system.The main work is as follows:1.Based on the domestic and foreign research background of personalized recommendation system,this paper introduces the theory and implementation process of recommendation algorithm based on association rules and collaborative filtering,as well as their respective advantages and disadvantages.2.Aiming at the problem of cold start of new users in collaborative filtering recommendation algorithm,the recommendation algorithm based on popularity is analyzed,and a popularity recommendation algorithm with user clustering is proposed.Through this method,the cold start problem of new registered users in collaborative filtering recommendation technology can be properly handled.3.According to the influence of the difference of popularity and popularity on similarity calculation in collaborative filtering recommendation,a collaborative filtering recommendation algorithm based on project(i.e.curriculum)popularity weighting and adaptive threshold is proposed.The algorithn relieves the influence of popularity on similarity calculation by setting adaptive popularity threshold and penalty weight,and introducing them into similarity calculation formula.4.Realize the design and development of recommendation system.The requirements of recommendation system are analyzed from both functional and non-functional aspects.Personalized...