Design And Implementation Of Audience Behavior Digital Service Platform In Wisdom Museum

With the development of information age,information overload problem has been become increasingly severely.To a certain extent,classification search and search engine solve the information overload problems,but instead different users get the same search result and don't reflect the users' personalized when they input the same keywords.Therefore personalized recommendation system came into being.Given the current development of the Wisdom Museum,for providing better personalized service for the audience,it is a relatively new research focus that the recommendation system technology joins the construction of Museum Audience Service.In this paper,based on the construction of Wisdom Museum Audience Service in Guangdong,at first,this paper introduces the subject background and current research of related technologies.And then it analyzes the purpose and significance in this paper.Secondly,it introduces the algorithm of the recommendation system and the mainstream computing platform for Big Data---Hadoop.And then Audience Behavior Digital Service Platform in Wisdom Museum is designed,including the platform about requirement analysis,overall architecture design,and various functional sub-module,etc.For design and implementation of the recommendation engine computing sub-module,the paper trys to combine Fuzzy K-Means clustering algorithm with collaborative filtering recommendation algorithm by analyzing and researching the related recommendation system algorithm.It accelerates the calculation of similarity with audience or exhibits and improves the operational efficiency by fuzzy clustering algorithm for parallel computing.By Hadoop application platform,it resolves to the problem about recommending the exhibits of target audience interested in.Hybrid recommendation algorithm in recommendation engine mainly includes: User-Based Collaborative Filtering Algorithm Based On Fuzzy Clustering Parallelized(UBCFFCP),Item-Based Collaborative Filtering Algorithm Based On Fuzzy Clustering Parallelized(IBCFFCP),and Collaborative Filtering On Dual Fuzzy Clustering Algorithm In Parallelization(CFDFCP).Finally,this paper makes experimental analysis for Audience Behavior Digital Service Platform and verify hybrid recommendation algorithm in the recommendation engine that has a better recommendation result by using Mahout tools.