Developing a maturity-based workflow for the implementation of ML-applications using the example of a demand forecast

The aim of the article is to present a guideline that has been developed in the form of a workflow to identify the capability of an organization to implement machine learning (ML) applications on the one hand and, on the other hand, to describe a maturity-dependent procedure for the development of an ML application based on this knowledge. With the help of the guideline, application-specific requirements can be identified based on the phases of the development process of an ML application adapted to the corporate environment. The article begins with the motivation for using machine learning methods and presents the challenges in implementing these methods. Based on a literature review, a maturity-based approach is designed and the developed and adapted development phases from the literature are described in a more detailed way. The individual characteristics of certain phases are specified based on the maturity level. As well, the weighting of certain maturity dimensions of the respective phase is highlighted. The article ends with an outlook on the further development of the created guideline.