Discovery of user preferred access patterns from web logs

Web users´ interest and preference can be revealed by such factors as access frequencies and time durations of web pages, in which time durations on a web page are denoted by a fuzzy linguistic variable. Thus, a novel concept of fuzzy preference considering the relative access frequency and time durations of a web page is proposed to be along with another concept of support considering the absolute access frequency of a web page to measure web user interest and preference. According to these two concepts, an algorithm is developed for mining user interesting/preferred access patterns from Frequent Link and Access Tree (FLaAT), which stores all user access information extracted from web logs.