Research On Fusion Recommendation Method Based On Content And Collaborative Filtering

Under the background of the rapid development of information technology at home and abroad,with the rapid popularization and application of mobile Internet and cloud computing,big data and other technologies,the Internet formally entered the era of Web3.0,making the global data volume show exponential growth.More and more information resources are filled with the network world and people's daily life,and the phenomenon of "information overload" appears,which makes it impossible for users to obtain accurate and effective information accurately.Recommendation system has been widely used in e-commerce,online social network,news media,video music and so on.Although there are some excellent solutions such as classified catalogs and search engines before this,because of the huge amount of data,the classified catalogs can not meet the requirements,and gradually withdraw from the view of people.Because of the characteristics of the search engine,the search engine has higher requirements for users' retrieval,discrimination ability and knowledge level,and it can only meet the needs of some users gradually.The birth of recommendation system can realize therecommendation of personalized content and fill in the field of personalized recommendation.Blank.The specific contents of this thesis are as follows:1.The recommendation system is summarized,and the present situation and existing problems of the recommendation system are analyzed.2.Clustering generates virtual user clusters,reduces the search range of nearest neighbors and improves the search rate,thus solving the time-consuming problem of traditional collaborative filtering algorithms.3.For the improved collaborative filtering algorithm,we continue to analyze the existing problems.Based on this,a new algorithm is proposed in this thesis.It is a recommendation method which combines content-based recommendation method and improved collaborative filtering algorithm.The feature of the project is extracted and the user score set combined with the feature is formed.Then the K-means cluster is used to realize the recommendation.The experimental results show that this method can solve the problem of data sparsity and cold start obviously.The experimental results show that the improved fusion algorithm runs faster than the traditional single algorithm.Moreover,the improved fusion algorithm can improve the data sparsity and cold start,and improve the quality of recommendation.The experimental results show that the improved fusion algorithm runs faster than the traditional single algorithm.Moreover,the improved fusion algorithm can improve the data sparsity and cold start,and improve the quality of recommendation.