An Efficient Method for Mining Clickstream Patterns

Recently, hybrid approaches, which combine an FP-tree-like data structure with an interaction-based approach, are efficient approaches for mining frequent itemsets. However, applying those approaches for sequential pattern mining arose some challenges. In this paper, we introduce a hybrid approach for a specific version of sequential pattern mining, clickstream pattern mining, with our proposed B-List structure and SMUB algorithm. The SMUB algorithm exploited the B-List structure that is generated from the SPPC tree and the B-List intersection are used to discover all sequential patterns in the given sequence database. Via our experiments on various databases, SMUB has been shown to be more efficient than the current state-of-the-art algorithm, CM-Spade, in terms of runtime, and scalability, especially on huge databases with very small thresholds.