Mining time constrained fuzzy association rules and sequential patterns from WEB logs

ABSTRACT MS Thesis MINING TIME CONSTRAINED FUZZY ASSOCIATION RULES AND SEQUENTIAL PATTERNS FROM WEB LOGS İsmail İŞERİ Fırat University Graduate School of Natural and Applied Sciences Department of Computer Engineering 2005, Page: 51 In this thesis, a novel web mining algorithm has been proposed to discover time constrained multiple-level browsing patterns from log data in web servers. Web mining of browsing including simple sequential patterns and association rules with browsing times has been studied recently. However, most of these works focus on mining browsing patterns of web pages directly. Moreover, almost all of them deal with single level rule and patterns. In this thesis, we have introduced the problem of mining browsing patterns on multiple levels of a taxonomy comprised of web pages. In addition, browsing time was considered and processed using fuzzy set concepts to form linguistic terms. The proposed algorithm thus discovers time constrained multiple-level relevant browsing behavior from linguistic data and promotes the discovery of coarsen granularity of web browsing patterns. The experimental results conducted on a real data set show the applicability and efficiency of the proposed method. Keywords: Data Mining, Web Usage Mining, Fuzzy Sets, Sequential Patterns, Association Rules VI