Design And Implementation Of Shopping Recommendation System Supporting High Concurrency

In recent years,with the rapid development of Internet technology,the Internet has become inseparable from people's lives.The development of e-commerce and express delivery industry has provided convenience for people to shop online,and directly changed people's shopping habits.The continuous innovation of information technology and the variety of products on the e-commerce platform have attracted a large number of customers,it provides opportunities and challenges for small and medium-sized e-commerce enterprises.Small and medium-sized e-commerce enterprises should pay more attention to the performance and services of their own online systems to improve the stability,availability,scalability,and concurrency of the system.Concurrency is one of the most important factors to improve the high performance of the system.When the number of users visiting the shopping system is too large,it will bring huge pressure to the server,which may cause the system response time to be slow or even the server crashes.It will reduce the viscosity of users,which may cause the loss of users of shopping platform.With the expansion of the scale of the shopping system and the development needs of enterprises,there will be more and more kinds of goods on the shopping platform,which will also cause many choices for platform users.Users may need to spend more time filtering out what they do not like.Therefore,the shopping platform should be able to provide users with high-quality recommendation services,so as to enhance the user 's shopping experience.It can not only attract more users,but also increase the sales of platform merchants,and it can well meet the functional needs of the shopping system.This thesis is dedicated to researching and implementing a shopping recommendation system that supports high concurrency for small and medium-sized e-commerce enterprises,it mainly improve the concurrency performance of the system and the real-time and accuracy of the recommended results.The specific work of the thesis is:(1)Using the micro service framework Spring Cloud combined with the SSM architecture as the infrastructure for the development of shopping recommendation systems,the shopping recommendation system has a portal system and a management system.(2)Through the research and design of distributed clusters,Nginx load balancing,Redis,Solr,MQ,database read-write separation and other technical solutions,it can improve the concurrent performance of different levels of the system.(3)Research the Hadoop platform.The personalized recommendation module utilizes the parallel computing capability of Map / Reduce and the distributed storage capability of HDFS to migrate the processing and computing work of commodity data information to be recommended,thus improving the recommendation real-time feedback speed.(4)This paper studies and analyzes the mainstream recommendation algorithm,and it adopts the collaborative filtering algorithm based on items as the recommendation algorithm of the shopping recommendation system,the recommendation algorithm was improved in combination with the user registration label to improve the recommendation accuracy,and gives the solution of cold start of user recommendation for different identities of the shopping recommendation system.