An end-to-end process model for supervised machine learning classification: from problem to deployment in information systems
Koehl, Niklas J.
Karlsruher Institut für Technologie (KIT)
Extracting meaningful knowledge from (big) data represents a key success factor in many industries today. Supervised machine learning (SML) has emerged as a popular technique to learn patterns in complex data sets and to identify hidden correlations. When this insight is turned into action, business value is created. However, common data mining processes are generally not tailored to SML. In addition, they fall short of providing an end-to-end view that not only supports building a ”one off” model, but also covers its operational deployment within an information system. In this research-in-progress work we apply a Design Science Research (DSR) approach to develop a SML process model artifact that comprises model initiation, error estimation and deployment. In a first cycle, we evaluate the artifact in an illustrative scenario to demonstrate suitability. The results encourage us to further refine the approach and to prepare evaluations in concrete use cases. Thus, we move towards contributing a general process model that supports the systematic design of machine learning solutions to turn insights into continuous action.
Data mining process , Supervised machine learning , Information systems
Hirt, R., Kuehl, N. J. and Satzger, G. (2017) 'An End-to-End Process Model for Supervised Machine Learning Classification: From Problem to Deployment in Information Systems'. In: Maedche, A., vom Brocke, J., Hevner, A. (eds.) Designing the Digital Transformation: DESRIST 2017 Research in Progress Proceedings of the 12th International Conference on Design Science Research in Information Systems and Technology. Karlsruhe, Germany. 30 May - 1 Jun. Karslruhe: Karlsruher Institut für Technologie (KIT), pp. 55-63