Mining individual access patterns by the time preference of a web user

Mining a web user´s access patterns can efficiently reveal the traits of personal behavior, e.g. the user´s interest and inherent surfing habits. According to a user´s access sequences which generate residence time and time intervals, a novel data structure called parent index tree of users´ frequent access patterns (UFAP-PIT) is developed for mining individual access patterns. Furthermore, the residence time of personal access pages in Web page classes which are extracted from UFAP-PIT is processed by K-means clustering. Meanwhile, the time interval between a pair of Web page classes is processed in the same way. Finally, our algorithm can effectively construct personalized user access patterns by the time preference. The experimentation has shown that the proposed algorithm presents a marked difference in personal behavior patterns by time attributes.