Loading...
Description
Enhancing data quality (DQ) to meet the requirements of digital processes is an essential prerequisite for any successful digitalization initiative. Machine Learning (ML) has become one of the most promising technological advances for enterprises to lower the barriers towards achieving required levels of DQ. Developing and deploying these ML based DQ systems in an organizational setting, however, is linked to a range of organizational and technical requirements and implications that researchers and enterprises have only started to comprehend. Based on an action design research approach, this study seeks to develop a comprehensive ML based solution for DQ controls, an essential instrument in Data Quality Management. Following the findings of the project, we propose a first set of design principles for ML based DQ controls as well as a set of differences to traditional DQ approaches.
Recommended Citation
Walter, Valerianne; Gyoery, Andreas; and Legner, Christine, "Deploying machine learning based data quality controls – Design principles and insights from the field" (2022). Wirtschaftsinformatik 2022 Proceedings. 4.
https://aisel.aisnet.org/wi2022/ai/ai/4
Deploying machine learning based data quality controls – Design principles and insights from the field
Enhancing data quality (DQ) to meet the requirements of digital processes is an essential prerequisite for any successful digitalization initiative. Machine Learning (ML) has become one of the most promising technological advances for enterprises to lower the barriers towards achieving required levels of DQ. Developing and deploying these ML based DQ systems in an organizational setting, however, is linked to a range of organizational and technical requirements and implications that researchers and enterprises have only started to comprehend. Based on an action design research approach, this study seeks to develop a comprehensive ML based solution for DQ controls, an essential instrument in Data Quality Management. Following the findings of the project, we propose a first set of design principles for ML based DQ controls as well as a set of differences to traditional DQ approaches.