nMITP-Miner: An Efficient Method for Mining Frequent Maximal Inter-transaction Patterns

In this paper, we introduce an algorithm for mining Frequent Maximal Inter-Transaction Patterns (FMITPs), named nMITP-Miner, utilizing the ITN-list structure. Our proposed algorithm offers several significant contributions. Initially, we implement efficient pruning strategies to promptly eliminate infrequent 1-patterns. Subsequently, we propose a strategy based on the ITN-list structure, aimed at minimizing the search space and rapidly identifying all FMITPs. Thereafter, our algorithm applies Depth First Search (DFS) to traverse and generate all FMITPs along with their corresponding ITN-lists. Finally, comprehensive experiments are conducted to validate the effectiveness and efficiency of nMITP-Miner compared to other existing methods, such as tMITP-Miner and dMITP-Miner, with a specific focus on runtime and memory usage.