Efficient Single-Pass Mining of Weighted Interesting Patterns

Mining weighted interesting patterns (WIP) [5] is an important research issue in data mining and knowledge discovery with broad applications. WIP can detect correlated patterns with a strong weight and/or support affinity. However, it still requires two database scans which are not applicable for efficient processing of the real-time data like data streams. In this paper, we propose a novel tree structure, called SPWIP-tree (Single-pass Weighted Interesting Pattern tree), that captures database information using a single-pass of database and provides efficient mining performance using a pattern growth mining approach. Extensive experimental results show that our approach outperforms the existing WIP algorithm. Moreover, it is very efficient and scalable for weighted interesting pattern mining with a single database scan.