The Development Of Personalized Recommendation System For "Yummy77" Website

The rapid development of Internet and electronic commerce has brought great convenience to people's daily life,and it has become more convenient and fast.But at the same time,there are a large number of commodity information will appear in the electronic commerce website.This,in turn,will users dazzled when shopping,helpless.How to let users from a large number of commodity information,to get the real useful part of their information,how to allow users to buy the real taste of their own goods.This needs personalized product recommendation system,which through the analysis of the user's behavior,you can guess the user's preferences,for the user's personal preferences,make personalized recommendation."Yummy 77"e-commerce sites are mainly engaged in the online sales of fresh products,the site has no personalized product recommendation system.But with the sale of the demand grow with each passing day,on the recommendation of personalized products is more and more urgent.According to this requirement,the design and implementation of personalized recommendation system based on commodity distributed collaborative filtering algorithm^[6].The main contents of this paper are as follows:In this paper,first,the application status of recommendation system in the domestic and foreign research and analysis,based on recommendation system definition,function,structure and classification of into the comb,and the mainstream of the current recommendation algorithm and recommendation system were studied,and their advantages and disadvantages are analyzed.Secondly,this paper on the status of operation and maintenance of delicious feast of e-commerce sites were analyzed,and combined with the site's own characteristics such as large number of registered users,massive user behavior data,order data source and diversification characteristics,delicious feast personalized recommendation system,carries on the demand analysis,architecture design,detailed design and implementation.According to the user behavior data,"Yummy 77"personalized recommendation system uses a Hadoop based distributed collaborative filtering algorithm.Effective solution for users of mass data uploading,similarity calculation,storage and transfer and other issues recommendation results,and realize the weak coupling between e-commerce sites and personalized recommendation system.At the end of the personalized recommendation system for performance testing and functional testing comprehensive,the test results show that the system functions correctly,to achieve the system requirements design requirements.