New Techniques for Data Preprocessing Based on Usage Logs for Efficient Web User Profiling at Client Side

Abstract

User profiling at client side is an important key task for intelligent information delivery in Web environment. However, several preprocessing also should be carried out prior to constructing user profiles. Moreover, the client side preprocessing should be carried out mostly based on usage logs without knowledge about contents and structures of Websites. In this perspective, we conducted an experiment in users’ natural Web environments during a period of several days, and analyzed collected usage logs and user feedbacks. Based on the results, we designed three important preprocessing techniques and evaluated the performances. We found that the required data preprocessing tasks can be conducted efficiently at client side based solely on usage logs.