Evaluating the benefits of empowering model‐driven development with a machine learning classifier. (6th August 2022)

• Record Type: Journal Article
• Authors: Marcén, Ana C.; Pérez, Francisca; Pastor, Óscar; Cetina, Carlos
• Is Part Of: Software, practice & experience. Volume 52:Number 11(2022)
• Page Start: 2439
• Page End: 2455
• Publication Date: 2022-08-06
• Subjects: data‐oriented software systems -- machine learning -- model‐driven development; Computer software -- Periodicals; Computer programming -- Periodicals; Computer programs -- Periodicals; 005.3
• DOI: 10.1002/spe.3133
• Languages: English
• ISSNs: 0038-0644

Abstract

Abstract: Increasingly, the model driven engineering (MDE) community is paying more attention to the techniques offered by the machine learning (ML) community. This has led to the application of ML techniques to MDE related tasks in hope of increasing the current benefits of MDE. Nevertheless, there is a lack of empirical experiments that evaluate the benefits that ML brings to MDE. In this work, we evaluate the benefits of empowering model engineers of model‐driven development (MDD) with an ML classifier. To do this, we tackled how to embed the ML classifier as part of the MDD. Then, this was evaluated using two different real industrial cases. Our results show that despite the ML part takes an extra effort, the use of the ML classifier pays off in terms of the quality results, the perceived usefulness, and intention to use.