An Efficient Algorithm for Maintaining Frequent Closed Itemsets over Data Stream

Data mining refers to the process of revealing unknown and potentially useful information from a large database. Frequent itemsets mining is one of the foundational problems in data mining, which is to discover the set of products that purchased frequently together by customers from a transaction database. However, there may be a large number of patterns generated from database, and many of them are redundant. Frequent closed itemset is a wellknown condensed representation of frequent itemset, and it provides complete information of frequent itemsets. Extensive studies have been proposed for mining frequent closed itemsets from transaction database, but most of them do not take streaming data into consideration. In this paper, we propose an efficient algorithm for maintaining frequent closed itemsets over data streams. Whenever a transaction is added to database, our approach incrementally updates the information of closed itemsets and outputs updated frequent closed itemsets based on user-specified thresholds. The experimental results show that our approach outperforms previous studies.