For joint work with colleagues from St. Petersburg, Russia, Michael Fellmann and Kurt Sandkuhl received the best paper award at the 14th IFIP WG 8.1 Working Conference on the Practice of Enterprise Modelling (PoEM 2021) in Riga, Latvia. The paper on „Machine Learning-Based Enterprise Modeling Assistance: Approach and Potentials“ was authored by Nikolay Shilov, Walaa Othman, Michael Fellmann and Kurt Sandkuhl. Nikolay Shilov is senior researcher at the St. Petersburg Federal Research Center of the Russian Academy of Sciences (SPC RAS); Walaa Othman a student member in this group. The paper is a result of the long-standing cooperation of the lab Computer-Aided Integrated System of Prof. Alexander Smirnov at SPC RAS and the Business Information Systems group at Rostock University of Prof. Kurt Sandkuhl.
The subject of the paper are concept and experimental testing of an assistive function for enterprise modeling. The motivation for this is first of all the realization that large model collections with thousands of enterprise architecture models exist in organizations. However, the knowledge contained in them is not yet systematically and explicitly reused in the construction of new models. One of the reasons for this is that knowledge of all relevant models requires a time-consuming review and familiarization with the model inventory, which is often difficult to implement in practice. The method described in this paper represents a possible solution to this problem by enabling the fully automatic generation of relevant suggestions during model construction via machine learning. By means of the suggestions, modeling can potentially be simplified and accelerated, and knowledge from a large number of existing models can be incorporated into the construction of new models.