An Efficient Web Mining Algorithm For Session Path Patterns

Abstract

Data mining in the World Wide Web has many practical applications. Using efficient web mining algorithms to discover WWW browsing path patterns can benefit web users as well as administrators. For example, effective information retrieval and better understanding of the user behaviors can help attract more visitors and improve service of E-commerce. However, d ifferent definitions of a browsing path result in mining different knowledge. We define a user session as a complete browsing path from the entry page to the last page a user visited. An efficient mining algorithm was proposed to find patterns of frequent session path. We introduce a new measurement of browser’s interest in a path as an inclination value. In this paper, we show that our implementation takes less memory space since the algorithm produces fewer candidate transactions than that of others. Scanning the sorted database only once, we can quickly find all large session paths regardless of the support value.