| With the rapid development of the Internet industry,Internet users are gradually increasing,which will inevitably lead to the rapid growth of data and information.It becomes difficult for users to find out their favorite information from a huge amount of information.Although an ordinary personalized recommendation system can make recommendations based on the historical behavior of users,there are still some setbacks such as data sparsity and cold start.In addition,with loads of schools of thought appearing in the e-commerce industry,there is also an urgent need to select the same products among various e-commerce websites.At the same time,with the rapid development of container technology,the development and deployment are fast and comvenient,which makes the technology prospect very bright.The following tasks have been completed:1)Personalized recommendation algorithm have been researched and analyzed,and some of the problems of this algorithm have been solved.Before the algorithm is implemented,the user clustering is performed to improve the recommendation accuracy and operational efficiency of the algorithm.2)The multi-electricity supplier's comprehensive recommendation system has been analyzed in many aspects and the system's three-tier architecture(interaction layer,business logic layer and database layer)has been designed and implemented.3)With Docker container technology and Nginx load balancing,the system is deployed on Docker clusters,which makes the system backward compatible and secure to a large extent.In this paper,user clustering algorithm and collaborative recommendation algorithm are combined to improve the recommended system accuracy and operational efficiency,while the Python-based Flask framework and Docker container technology are used to meet the rapid development of the system deployment,backward compatibility and reliability.After testing,the system can meet the needs of e-commerce sites and recommender systems. |
PDF Full Text Request