Mining frequent closed sequential patterns with non-user-defined gap constraints

Frequent closed sequential pattern mining plays an important role in sequence data mining and has a wide range of applications in real life, such as protein sequence analysis, financial data investigation, and user behavior prediction. In previous studies, a user predefined gap constraint is considered in frequent closed sequential pattern mining as a parameter. However, it is difficult for users, who are lacking sufficient priori knowledge, to set suitable gap constraints. Furthermore, different gap constraints may lead to different results, and some useful patterns may be missed if the gap constraint is chosen inappropriately. To deal with this, we present a novel problem of mining frequent closed sequential patterns with non-user-defined gap constraints. In addition, we propose an efficient algorithm to find the frequent closed sequential patterns with the most suitable gap constraints. Our empirical study on protein data sets demonstrates that our algorithm is effective and efficient.