Abstract: In order to improve the efficiency and performance of mass data mining for FLWAP-mine algorithm, based on reducing the time of scanning the database and the Apriori properties of sequential pattern, the FLWAP-mine algorithm is improved. It constructs the projection tree by scanning the database once, and then deletes the non-frequent events from the projection tree by pruning. The improved FLWAP-mine algorithm adopts the idea of the projection tree to search the projection tree of the current pattern, reduces the search area by pruning the projection tree. The experimental results show that the improved FLWAP-mine algorithm performs better than the previous one when the data set is large or the minimum support threshold is small.

Key words: sequential pattern, FLWAP-mine, projection tree, pruning