Author(s)

J. D. Velàsquez, H. Yasuda, T. Aoki & R. Weber

Abstract

A generic Data Mart architecture to support Web mining J. D. Ve15squez1, H. Yasudal, T. Aokil & R. Weber2 ' ~esearch Center for Advanced Science and Technology, University of Tokyo 2~epartment of Industrial Engineering, University of Chile Abstract Visits in a Web site leave behind important information about the behavior of the visitors. This information is stored in log files, which can contain many registers but part of them do not contain relevant information. In such cases, user behavior analysis turns out to be a complex and time-consuming task. In order to analyze Web site visits, the relevant information has to be filtered and studied in an efficient way. We introduce a generic Data Mart architecture to support advanced Web mining, which is based on a Star model and contains the relevant historical data from visits to the Web site. Its fact table contains various additive measures that support the intended data mining tasks, whereas the dimension tables store the parameters necessary fo

Keywords