Modeling machine learning requirements from three perspectives: a case report from the healthcare domain

Abstract

Implementing machine learning in an enterprise involves tackling a wide range of complexities with respect to requirements elicitation, design, development, and deployment of such solutions. Despite the necessity and relevance of requirements engineering approaches to the process, not much research has been done in this area. This paper employs a case study method to evaluate the expressiveness and usefulness of GR4ML, a conceptual modeling framework for requirements elicitation, design, and development of machine learning solutions. Our results confirm that the framework includes an adequate set of concepts for expressing machine learning requirements and solution design. The case study also demonstrates that the framework can be useful in machine learning projects by revealing new requirements that would have been missed without using the framework, as well as, by facilitating communication among project team members of different roles and backgrounds. Feedback from study participants and areas of improvement to the framework are also discussed.