PROWL: An Efficient Frequent Continuity Mining Algorithm on Event Sequences

摘要

Mining association rule in event sequences is an important data mining problem with many applications. Most of previous studies on association rules are on mining intra-transaction association, which consider only relationship among the item in the same transaction. However, intra-transaction association rules are not a suitable for trend prediction. Therefore, inter-transaction association is introduced, which consider the relationship among itemset of multiple time instants. In this paper, we present PROWL, an efficient algorithm for mining inter-transaction rules. By using projected window method and depth first enumeration approach, we can discover all frequent patterns quickly. Finally, an extensive experimental evaluation on a number of real and synthetic database shows that PROWL significantly outperforms previous method.