Set-based approach in mining sequential patterns

In this paper, we describe a set-based approach for mining association rules and finding frequent sequential patterns in customer transactional databases. The set-based approach is a direct improvement of the original association rule mining algorithms proposed by R. Agrawal and R. Skrikant. Our approach relaxes the constraints described in Apriori (All/Some), and improves the performance while being more user-oriented and self-adaptive than the probabilistic knowledge representation. We compare the performance of the improved algorithms with results from an experimental study. The approach can be extended to more set-based mathematical models for further data analysis in order to discover hidden knowledge and patterns with the improved workflow and set-based representation.