Efficient Mining of Sequential Patterns in a Sequence Database with Weight Constraint

Abstract

Sequence pattern mining is one of the essential data mining tasks with broad applications. Many sequence mining algorithms have been developed to find a set of frequent sub-sequences satisfying the support threshold in a sequence database. The main problem in most of these algorithms is they generate huge number of sequential patterns when the support threshold is low and all the sequence patterns are treated uniformly while real sequential patterns have different importance. In this paper, we propose an algorithm which aims to find more interesting sequential patterns, considering the different significance of each data element in a sequence database. Unlike the conventional weighted sequential pattern mining, where the weights of items are preassigned according to the priority or importance, in our approach the weights are set according to the real data and during the mining process not only the supports but also weights of patterns are considered. The experimental results show that the algorithm is efficient and effective in generating more interesting patterns.