With progressive expansion in size and complexity of Web sites on WWW, much research has been done on the discovery of useful user traversal patterns. Most existing approaches focus on finding Web association rules, traversal paths or sequential patterns from Web logs. In this paper, we present a new pattern, Web traversal walk, for analysis of the structural navigation activities of Web users. A Web traversal walk is a structural sequence of forward and backward traversal paths. An efficient algorithm, Fast-Walk, is proposed to discover the Web traversal walks. In Fast-Walk, a tree structure is constructed in memory from Web logs and the frequent Web traversal walks are generated from the tree structure. Experimental results show the efficiency and scalability of Fast-Walk.

Date Published: 12 March 2002
PDF: 12 pages
Proc. SPIE 4730, Data Mining and Knowledge Discovery: Theory, Tools, and Technology IV, (12 March 2002); doi: 10.1117/12.460237

Published in SPIE Proceedings Vol. 4730:
Data Mining and Knowledge Discovery: Theory, Tools, and Technology IV
Belur V. Dasarathy, Editor(s)