An Efficient Algorithm for Mining Frequent Closed Inter- Transaction Patterns
Thanh-Ngo Nguyen, Loan T. T. Nguyen, Bay Vo, Ngoc Thanh Nguyên · 2019
One of the key tasks of data mining is frequent pattern mining (FPM), discovering patterns that frequently occur in transactions of transaction databases. A majority of methods of FPM only mine patterns in the same transactions, they do not consider patterns across several transactions (intertransaction patterns- ITPs) in a transaction database. So far, a number of algorithms have been proposed to mine ITPs. However, they still have many problems that need improving to make them more efficient in mining time and memory usage. This paper presents an efficient method of mining frequent closed inter-transaction patterns (FCITP) from a transaction database. Our proposed method generates inter-transaction patterns at the 1-pattern level. Besides, we propose an efficient pruning strategic to prune infrequent inter-transaction 1- patterns and quickly generate frequent inter-transaction 2- patterns. To eliminate non-closed patterns, the proposed method adopts properties of frequent closed patterns and pruning strategies. In addition, a Diffset-based strategy for the efficient computation of the supports of closed inter-transaction patterns is described. Experiment results conducted on several datasets of various characteristics show that our algorithm outperforms the state-of-the-art algorithms in terms of mining time and memory usage in most cases.