An Efficient Approach to Discovering Repeated Sequential Patterns in Large Transaction Databases

Abstract

Data mining, also known as knowledge discovery, is recognized as an area of promising research. It can be defined as the efficient discovery of interesting patterns from a large database. Mining sequential patterns involves discovering patterns in sequences of events with a user-defined minimum support. The support of a pattern is the number of data-sequences that include the pattern. The previous approach, the GSP（generalized sequential patterns）algorithm, checks whether a data-sequence contains a specific sequence for counting candidates requiring alternates between the forward and backward phases until all the items are found. The higher complexity of operations is performed repeatedly to detect certain particular items. In this paper, a DRSP algorithm using two additional Order list and Check list data structures is proposed, to reduce the complexity of sequence search. The proposed approach can also discover the repeated sequential pattern, which include several overlapped items. Experimental results indicate that the proposed searching approach outperforms GSP algorithm.