Research On Personalized Recommendation Algorithm Based On User Interest Model

Usually,the recommendation process of personalized recommendation system is carried out in stages.The first stage is interest modeling,the second stage is project matching,and the last stage is recommendation result feedback.The first stage is mainly to obtain users' interests and needs,and then generate a user model including users' preferences,backgrounds and needs.The second stage is to match resources and users' interests by using different recommendation algorithms based on the user model acquired in the first stage.The final stage is to generate a list of recommendations to be presented to the user.As the basis of personalized recommendation system,the construction of user model is very important.Whether the user's interests and preferences can be accurately obtained determines the effect of personalized recommendation.In view of the shortcomings of the existing work in user interest modeling using resource feature selection,the data source is single and the user's real interest can not be captured.This paper focuses on the construction of interest model,and then designs an efficient recommendation algorithm under the accurate user interest model.The main contents of this paper are as follows:Firstly,the current research status of user interest modeling at home and abroad is summarized,and the basic algorithms and common algorithms in the field of recommendation system are introduced.The related research and application of group wisdom and representation learning technology are analyzed.Secondly,this paper proposes a user interest modeling method based on group wisdom,which integrates the user's browsing,evaluation and other behaviors,and uses group wisdom and representation learning technology to achieve accurate multi-source data modeling of user interest,in order to improve the effect of personalized recommendation.Thirdly,an algorithm based on user interest model is proposed.Based on the user's historical interest model,the user's interest model is obtained by combining theuser's behavior information.Finally,the similarity between candidate items and users is calculated,and TOP-N recommendation is made according to the similarity calculation results.Finally,in the experiment,this paper takes the movie recommendation task as an example to do empirical research,and verifies the effectiveness of the proposed method on real data.The experimental results show that,compared with the traditional user interest modeling method,the user interest modeling method based on group wisdom takes massive data as the research basis,can more accurately depict user interest,and thus achieve better recommendation effect.