Research For Electronic Commerce Pattern Based On Data Mining

With the rapid growth and the popularization in Internet and WWW, Electronic Commerce has caught more attention from researchers. They want to take the advantage of the new commerce to gain more profit. Web has already become the basis of the Electronic Commerce. By applying the approaches of Data Mining into the Electronic Commerce, the owner of the Electronic Commerce can find out the really useful knowledge from the mass of data to make a right decision. Web Data Mining has been combined with the Electronic Commerce on this occasion. It is a new branch of Data Mining.The objects of Web Mining include all kinds of Web data: content of Web pages, structure between pages, and usage information of users. The Web Usage Mining mines Web log records to discover users access patterns of Web pages. Analyzing and exploring regularities in Web log records can identify potential customers for Electronic Commerce, enhance the quality and delivery of Internet information services to the end users, and improve Web server system performance.The thesis addresses the research of the Web Usage Mining .By mining the Web log records, we can obtain the knowledge about user access manners, which can find out the user access pattern and can perform personalization Web pages recommendation. Therefore there is a full solution for the Web site of the Electronic Commerce.The main work of the thesis is as follow.1. The thesis outlines the current research status at home and abroad, introduces the significance of the thesis and some relative theories.2. A data preprocessing method has been used. First, the Web log file must be analyzed. The raw Weblog data need to be cleaned, condensed and transformed in order to retrieve and analyze significant and useful information. Then the Web log records are organized into some transactions or sessions in a database.3. The maximum forward path method can transform the data into transactional databases, which are appropriated for mining.4. Association rules mining can be used in order to find out the user access pattern in Electronic Commerce. Since it can not really fit in the Web log mining, the traditional association rules mining must be improved. The thesis presents a new frequent path algorithm, which can find out the user access pattern in Electronic Commerce. These models can help the owner of Electronic Commerce to improve the Web site designing.5. The thesis presents a new collaborative filtering algorithm, which can classify the users in Electronic Commerce, can perform personalization Web pages recommendation and can provide individual service.Finally, this thesis summarizes the author's works and discusses the future works.