Maintenance Algorithm of Sequential Patterns Based on Implicit Segmentation

Constraints are essential for many sequential pattern mining applications. Duration constraint defined the time-stamp difference between the first and last transactions must be longer or shorter than a given integer. But it can not solve the problem that time-stamp difference between the first and last transactions is infinite. In this paper a new constraint called the partial-order constraint is introduced, which can solve the problem of time-stamp difference between the first and last transactions is infinite. The algorithm of simple updating frequent sequential pattern and sequential pattern maintenance algorithm are proposed based on partial-order constraint. The first algorithm is very simple and suits for the case that the constraint does not change the number of customer sequences. The second algorithm makes full use of the information obtained from previous mining processes, and adopts implicit segmentation technique which segments dissatisfy constraint sequences from the existing ones if their customer ids are the same.