Towards CRISP-ML(Q): A Machine Learning Process Model with Quality Assurance Methodology

Machine learning is an established and frequently used technique in industry academia, but a standard process model to improve success efficiency of machine applications still missing. Project organizations practitioners face manifold challenges risks when developing have need for guidance meet business expectations. This paper therefore proposes the development applications, covering six phases from defining scope maintaining deployed application. Business data understanding are executed simultaneously first phase, as both considerable impact on feasibility project. The next comprised preparation, modeling, evaluation, deployment. Special focus applied last running changing real-time environments requires close monitoring maintenance reduce risk performance degradation over time. With each task process, this work quality assurance methodology that suitable address identified form risks. drawn practical experience scientific literature, has proven be general stable. expands CRISP-DM, mining enjoys strong support, fails specific tasks. presented industry- application-neutral tailored with technical tasks assurance.