Research And Implementation Of Museum Collection Recommendation Algorithm Based On Audience Behavior

The rapid development of the Internet today generates a large amount of data in all walks of life.These data either explicitly or implicitly record the user's behavior and preferences,and use this data to discover many important information of the user.In recent years,the digitalization and intelligence of traditional museums have provided conditions for better service audiences,and museums can use the behavioral information of the audience to provide personalized services.Therefore,making full use of cloud computing,big data and recommendation systems to help viewers find collections that may be of interest to them,and to find people with similar interests,can provide a better experience for the audience.In this paper,based on the cold start and data sparsity problems faced by the recommendation algorithm,a content-based recommendation algorithm and a cluster-based recommendation algorithm are used for recommendation.This paper first introduces the research background of this topic,the current development status of the wisdom museum at home and abroad,and the domestic and international research of the recommendation system.At the same time,it expounds the theoretical basis of the personalized recommendation system,and the commonly used recommendation algorithm,evaluation index and similarity.The calculation,clustering algorithm and Hadoop big data platform are described,and their advantages and disadvantages are also analyzed.Secondly,this paper preprocesses the museum's collection data and audience data,and obtains three main attribute feature matrices.Then the initial clustering center of the fuzzy clustering algorithm is optimized,and the optimized algorithm is parallelized by MapReduce.The new MRKPFCM algorithm is obtained.The effectiveness of the algorithm is verified by experiments.Then,in order to solve the sparsity problem of collaborative filtering algorithm,the MRKPFCM algorithm is combined with the classical collaborative filtering algorithm.A parallel collaborative collaborative filtering recommendation algorithm based on MRKPFCM and audience and a parallel collaborative filtering recommendation algorithm based on MRKPFCM and collection are proposed.In addition,the above two recommended algorithms are weighted and mixed,and supplemented by the content-based recommendation algorithm to produce the final recommendation result.Finally,this paper builds a Hadoop-based museum collection recommendation system,which realizes the functions of viewing collection information,collection recommendation and recommendation system parameter adjustment,and uses traditional recommendation algorithm and MRKPFCM-based recommendation algorithm to carry out comparative experiments to verify the improved version.The algorithm improves the recommendation efficiency and accuracy of the recommendation algorithm.