Study On Technology Of Goods Personalized Push In E-commerce Based On Data Mining

At present, network technology and information technology develops rapidly, and in front of people is a sea of digital information resources, but everyone have their own specifical require for information. In traditional Web services, the information is distributed in the different levels of page, so people have to look out hard to find the information that they wanted from varieties of information. It wastes much of time and energy. Especially currently network economy surging, to meet the information needs of the user-specific personalization and save time for user judging, how to achieve a true sense of initiative, interactive, personalized service will be indispensable market methods in today’s businesses.Using Web mining, can not only get a lot of variety of seemingly unrelated Web data associated with them to extract useful knowledge that we need to, but also get the universal knowledge about the behavior and the way groups of users to access. Through the understanding and analysis of these knowledge such as user access behavior, frequency content act, we can extract the characteristics of the user, thereby personalize interface for user customization. The application of data mining technology can help to understand the meaning user interaction data containing, so the technology can be widely used in personalized push services in areas such as information intelligent service.In this thesis, I describe e-commerce personalized recommendation technology. I analyze the e-commerce technology and algorithm of the recommendation system deeply, summarize the status of e-commerce recommendation system, and build e-commerce recommendation system architecture. In the primise of meeting the system the recommended accuracy and real-time need the system settings recommended for offline mining and online in two parts, and realize that every part of the functional and technical. To meet the accuracy requirements of the e-commerce personalized recommendation, I make further research on clustering algorithm, merge ant colony algorithm and K-Means clustering algorithm, and design a new recommendation system. At last, I make experiments on the new recommendation system, and make summary of the experimental results and outlook.