Incremental mining of closed inter-transaction itemsets over data stream sliding windows
Journal of Information Science
| Hi-index | 0.00 |
Intertransaction frequent itemsets break the barriers of transactions, and extend the traditional single-dimensional intratransaction association rules to multidimensional intertransaction association rules. However the amount of intertransaction frequent itemsets becomes very large with the increase of the sliding window. Frequent closed itemsets can uniquely determine the set of all frequent itemsets and their exact frequency while they are far smaller than all frequent itemsets. In this paper, we introduce the notion of intertransaction frequent closed itemset, analyze its properties, and develop an efficient algorithm, IFCIA (Intertransaction Frequent Closed Itemsets Algorithm). The algorithm adopts division-based method and condition database to avoid generating large extended database, and uses bitmap structure and extended bitwise operations to generate candidate itemsets and count the support quickly. Experiments on real and synthetic databases show that IFCIA is an effective algorithm for mining intertransaction frequent closed itemsets.