Research And Implementation Of Personalized Intelligent Business Recommendation System Based On Big Data

With the continuous development and wide application of Internet technology,people’s life and work are accelerating while the pressure is increasing.In order to reduce the pressure of life,o2 o business model is born.The so-called o2 o refers to the combination of offline transactions and the Internet,which greatly facilitates people’s lives.However,at the same time,the information generated under this business model is growing exponentially and explosively,resulting in the problem of information overload.It is very urgent to solve the problem of information overload and reduce the time for users to search information.For this reason,the recommender system is born.It can accurately and effectively predict user preferences through the use of modern technology and algorithms.At present,the development of recommendation system is relatively mature,and it has been widely used in music,shopping,movies and other platforms,but it still has some problems,such as low recommendation efficiency,new users cold start and so on.In order to solve these problems,this thesis proposes a personalized intelligent business recommendation system based on big data platform.The research work of this paper mainly includes:(1)The similarity calculation that incorporates user attribute characteristics.The traditional similarity calculation method usually only considers the user’s rating data in the system,and ignores the similarity of user attributes.When a new user enters the system for the first time,there is a lack of rating data,and the system cannot recommend it,and there is a cold start of the new user problem.Among the user attributes,the attributes that most affect user interest are age and gender.Therefore,this study integrates user age and gender into the traditional similarity calculation,and can effectively recommend new users when they enter the system for the first time to solve the cold start problem.(2)Build a hybrid clustering model.Traditional collaborative filtering recommendation algorithms have a long running time and low recommendation efficiency when the amount of data is large.Therefore,in this paper,a hybrid clustering model(K-means+Canopy clustering model)is used to cluster users with high similarity together to form multiple clusters.When recommending,only need to be in the same cluster.Find the user’s nearest neighbors in the cluster,so as to shorten the search time,shorten the running time,and improve the recommendation efficiency..(3)Visualize recommendation results.Build a big data cluster,and perform data storage and operations on the cluster.At the same time,in order to make the research results more meaningful,the SSM(Spring +Spring MVC + My Batis)framework was used to build a system test platform to visualize the results.