Research And Application Of Recommender System Based On MongoDB

With the rapid and vigorous develop of electronic commerce, the way of our life ischanging unprecedentedly. The person who provides the data resource makes the directconnect with users，which offers more choices for the development of commerce model.Meanwhile, the resource is changing larger and larger, which make users get the effectiveinformation about all of the data resources become more difficult than before. Thetraditional research, which based on the key words, is different to research the target in themass data and multi-format data. How to get the effective information and how to dealwith the mass data are the problems of the mass data times. In this context, we hope thesystem to analysis user behavior data automatically and commend user relevant resource.The first part introduces the status of research and application of MongoDB at home andaboard, which describes the importance and superiority of application of commended model inthe big data system, simply. Then the second part emphasizes the status of application andresearch of the key technologies, which includes the application of commended system, theresearch of commended model and strategy, the development of NoSQL database, likeMongoDB, and even the data mining with Long Tail. The third part introduces the analysis ofthe design process of this system mainly, including requirements analysis, outline design,detailed design and database design and so on. The fourth part describes the achievement of it,including the software and hardware of the system development, core function and procedureand the achievement of distributed computing. The fifth part includes outline test, detailed test,the status of data and the performance analysis. The last part of the article offers the summariesand prospects about the further research.During the development and research process of this system, with the commendedmodel, it excavates the potency of data on the Long Tail. Meanwhile, use the distributedcomputing in the user behavior analysis which will improve the efficiency of processingdata and save the computing resources. With the test status, the system takes the SlopeOne algorithm in big data, which shows the superiority of prediction. And it could solvethe cold start in the commended system. With the NoSQL database as data store, itimprove the expansibility of data. The commended model and strategy also could be commended to other fields, and has a good promotion prospects.