iWAP: ASingle Pass Approach for Web Access Sequential Pattern Mining

Abstract

With the explosive growth of data availability on the World Wide Web, web usage mining becomes very essential for improving designs of websites, analyzing system performance as well as network communications, understanding user reaction, motivation and building adaptive websites. Web Access Pattern mining (WAP-mine) is a sequential pattern mining technique for discovering frequent web log access sequences. It first stores the frequent part of original web access sequence database on a prefix tree called WAP-tree and mines the frequent sequences from that tree according to a user given minimum support threshold. Therefore, this method is not applicable for incremental and interactive mining. In this paper, we propose an algorithm, improved Web Access Pattern (iWAP) mining, to find web access patterns from web logs more efficiently than the WAP-mine algorithm. Our proposed approach can discover all web access sequential patterns with a single pass of web log databases. Moreover, it is applicable for interactive and incremental mining which are not provided by the earlier one. The experimental and performance studies show that the proposed algorithm is in general an order of magnitude faster than the existing WAP-mine algorithm.