Discovering exclusive patterns in frequent sequences

This paper presents a new concept for pattern discovery in frequent sequences with potentially interesting applications. Based on data mining, the approach aims to discover exclusive sequential patterns (ESPs) by checking the relative exclusion of patterns across data sequences. ESP mining pursues the post-processing of sequential patterns and augments existing work on structural relations patterns mining. A three phase ESP mining method is proposed together with component algorithms, where a running worked example explains the process. Experiments are performed on real-world and synthetic datasets which showcase the results of ESP mining and demonstrate its effectiveness, illuminating the theories developed. An outline case study in workflow modelling gives some insight into future applicability.