Analysis And Display Of The User Trading Behavior In Modern Barter Industry

With the rapid development of Internet, many traditional industries have gradually gone to the Internet, and have another high speed development. The modern barter industry become popular because the use of the Internet platform and the developed logistics system. It greatly improves the flexibility and operability of the traditional barter, and solve the problems of today's inventory of Chinese enterprises. The new trading patterns and the general e-business models have similar points, but its unique way of doing business, will certainly bring new features.For e-commerce sites, the analysis of user behavior has been a popular field of study. The results based on user behavior analysis to improve the site structure, provide a more reasonable services to increase customer satisfaction, revenue for the site. In modern barter trade this new model, its analysis of user behavior, is still relatively scarce. Therefore, research in this area is very meaningful.This study is based on the Huamao electronic barter website, "Yitong website". On the site, users can operate within the system to barter, their behavior will be recorded to the operation of the service log file. Through the processing and analysis of log files to get the user's behavior patterns, and to visually display the results to the user, this paper's major work is as follows:(1) Described and analyz the Web log mining data preprocessing techniques, including data cleaning, session identification, path congruent fill key technology. For association rule mining algorithms for key studies, detailed description of the principle of Apriori algorithm and data mining process, analyze strengths and weaknesses of the algorithm and the algorithm for the lack of an improved algorithm to improve mining efficiency and accuracy. Through the test proved that the proposed improved algorithm is effective and reliable.(2) Describe the Yitong website function module design framework and the details of the transaction flow analysis of the system and operation mode. Details of the improved network mining algorithm is implemented in the Yitong website, and its mining results in the visual display system. System implementation using J2EE for development, with the mainstream open-source Java framework Struts+Spring+Hibernate. Implementation of data visualization using a free Javascript component Fusion Charts, its good composition function will show mining results are simple and intuitive to the user.The proposed algorithm in the Yitong website has been well achieved, and dig out a meaningful association rules, analysis of user behavior patterns, to improve the site to provide an effective reference for business decision support.