Study On E-commerce's Personalization Service Based On Web Usage Mining

With the development of communication, database, safety, data processing and Internet, E-commerce technology is becoming mature. At present, large-scale E-commerce enterprises such as taobao, ebay, alibaba and joyo are developing at unbelievable speed; other enterprises that don't engage in E-commerce are putting their goods to the net. It's obvious that traditional economic mode is being pursued and surpassed by network economy. Under traditional economic mode, enterprises provide personalized products and services to survive and succeed in competition. It is the same when the economic mode turns to be network. However, how do E-commerce websites provide personalized services and products? How to make the personalization more timely and accurate? All of these questions are main contents in this article.Chapter 1 firstly introduces the definition and the relationships of E-commerce's personalization and personalized E-commerce. Then it mentions that E-commerce should have personalized working mode from economic environment of E-commerce, the need of website strategy and working mode of enterprise. What's more, this chapter narrates the definition and processes of Web usage mining (Wum). At last, it lists the applications of Wum in E-commerceChapter 2 brings forward channels which are server, client, intermediary part and examples of data collecting. Finally it lists some common problems existed in these channels.Chapter 3 introduces the processes of data preprocessing, they are data cleaning, user identification, user session identification, path supplement and data formatting. Then every process is expatiated, some algorithms are given, some more effective methods are put forward and some related flow charts are appended.Chapter 4 narrates briefly the definition and classification of collaborative filter, introduces its processes in detail, and puts forward its defects. It brings forward k-means that is based on dynamic clustering and an example about it; the iterative results are depicted by charts, too.Chapter 5 mainly discusses the implementation of the E-commerce's personalized website based on Wum. It firstly puts forward the structure and functional orientation of recommendation engine. In addition, it sets up the system framework of the personalized mode in E-commerce from 3 layers. Following it introduces the website servers (Apache Tomcat Web server and Sqlserver2000 server) and creates the structure of log database. At last, it narrates the necessary functional modes and the improved personalizing strategies of E-commerce website.