Using a projection-based approach to mine frequent inter-transaction patterns

作者：Wang Chun Sheng; Chu Kuo Chung^*

来源：Expert Systems with Applications, 2011, 38(9): 11024-11031.

DOI：10.1016/j.eswa.2011.02.146

• Data mining
• Frequent pattern
• Inter-transaction pattern
• Projected database

In this paper, we propose an algorithm called PITP-Miner that utilizes a projection based approach to mine frequent inter-transaction patterns efficiently. The algorithm only searches for local frequent items in a projected database that stores potential local inter-transaction items and partitions the database into a set of smaller databases recursively. In addition, two pruning strategies are designed to further condense the partitioned databases and thus accelerate the algorithm. Our experiment results demonstrate that the proposed PITP-Miner algorithm outperforms the ITP-Miner and FITI algorithms in most cases.

• 出版日期2011-9