Context-sensitive textual recommendations for incomplete process model elements

Abstract

Many organizations manage repositories of several thousand process models. It has been observed that a lot of these models have quality issues. For the model collections we have worked with, we found that every third model contains elements with incomplete element names. While prior research has proposed techniques to close gaps on the structural level, approaches that address the naming of incompletely specified model elements are missing. In this paper, we propose three strategies for naming process elements and a context-sensitive ranking to present the most relevant naming recommendations to the user. We prototypically implemented our approach and conducted an extensive user experiment with real-world process models in order to assess the usefulness of the recommendations. The results show that our approach fulfills its purpose and creates meaningful recommendations.