Automatic extraction of process categories from process model collections

Many organizations build up their business process management activities in an incremental way. As a result, there is no overarching structure defined at the beginning. However, as business process modeling initiatives often yield hundreds to thousands of process models, there is a growing need for such a structure. This challenge calls for a technique to extract process categories from a set of process models automatically which yields an up-to-date view of the structure of a collection of process models. It also provides a means to check whether pre-defined process categories are still reasonable. In this paper, we introduce a technique for automatically extracting process categories from process model collections and test it using a collection from industry. The results demonstrate the usefulness of the technique by revealing issues of the pre-existing process categories. In this way, we contribute to the field of process model management and quality assurance.