Patterns Mining in Sequences Using Constraints

Data mining is a set of methods used in the process of KDD ( Knowledge Discovery in Data) in order to distinguish relationships and unknown patterns in the data. Mining patterns is an interesting technique and is widely used in data mining; its objective is to find the patterns that appear frequently in a database. The sequence mining is the pattern discovery problem in the sequences. A declarative approach has been proposed to solve this problem in order to transform this problem to an SAT model. In this paper, we propose a CSP-based encoding for the problem of discovering frequents and closed patterns in a sequence. We show that is possible to employee constraint programming techniques for modeling and solving a wide variety of constraint-based item-set mining tasks, such as frequent, closed and maximal. Preliminary experiments show that the new formulation is competitive and can outperform the SAT based approach on the considered sequences