Study On E-Commerce Site Structure Optimization Technologies

With the rapid increase of scope of E-commerce websites, users can not rapidly find interesting information in a large mount of data. Accordingly, how to improve the performance of websites and provide personalized convenient services is the key of the market competition of E-commerce websites. Website structure optimizing technique capable of dynamically adjusting website structure and satisfy user demands becomes research focus. Firstly, the thesis researches techniques of e-commerce, web data mining and website structure optimizion, mainly researches log data preprocessing and website structure optimizing method in website structure optimization, and analyzes the existing disadvantages of these techniques in detail.Secondly, considering the problem that the existing log data preprocessing techniques and path complementing algorithms can not solve processing of log data in dynamic website structure, this thesis presents the concept, construction method and graphic structure storing strategy of dynamic website structure. On this basis, the thesis improves the existing data preprocessing models and presents a path complementing algorithm based on page types in dynamic website structure.Thirdly, because the existing website structure optimizing techniques based on web mining have disadvantages in determining the content and duration of pages, and neglet the effect of the relation between page weight and visiting sequence in the curren session on the interest of users, this thesis uses association rule methods in web mining to mine web logs, and presents a website structure optimizing method based on dynamic link. By determing the content and duration of pages and normalizing the current session of websites visited by users, this method finds the interest of users. Tests prove that this method can increase the interest of users in links added dynamically and realize website structure optimization.Finally, to verify the path complementing algorithm and website structure optimizing method proposed, this thesis uses logs of Global Fashion Net as data sources and realizes this proposed algorithm with Java. The test results of preprocessing of data in dynamic website structure and the website structure optimizing method achieve expected objectives.