Mining Contrast Sequential Patterns with ASP

In this paper we address an extension of the sequential pattern mining problem which aims at detecting the significant differences between frequent sequences with respect to given classes. The resulting problem is known as contrast sequential pattern mining, since it merges the two notions of sequential pattern and contrast pattern. For this problem we present a declarative approach based on Answer Set Programming (ASP). The efficiency and the scalability of the ASP encoding are evaluated on two publicly available datasets, iPRG and UNIX User, by varying parameters, also in comparison with a hybrid ASP-based approach.