Recent studies on Machine learning (ML) and its industrial applications report that ML-enabled systems may be at a high risk of failure or they can easily fall short of business objectives. Cutting-edge developments in this field have increased complexity and also brought new challenges for enterprise information integration. This situation can even get worse when considering the vital importance of ML-enabled healthcare information systems (HEIS). Therefore, the main argument of this paper is that we need to adopt the principles of Industrial Information Integration Engineering (IIIE) for the design, development, and deployment processes of ML-enabled systems. A mixed research paradigm is adopted, and therefore, this study is conducted by following the guidelines and principles of Action Research, Design Science Research, and IIIE. The contributions of this study are two-fold: (a) to draw researchers’ and practitioners’ attention to the integration problems of ML-enabled systems and discuss them in view of IIIE, and (b) to propose an enterprise integration architecture for ML-enabled HEIS of a university hospital, which is designed and developed by following the guidelines and principles of IIIE.

最近关于机器学习 (ML) 及其工业应用的研究报告称，支持 ML 的系统可能存在很高的故障风险，或者它们很容易达不到业务目标。该领域的前沿发展增加了复杂性，也为企业信息集成带来了新的挑战。考虑到支持 ML 的医疗保健信息系统 (HEIS) 的重要性时，这种情况甚至会变得更糟。因此，本文的主要论点是我们需要采用工业信息集成工程（IIIE）的原则来设计、开发和部署支持机器学习的系统。采用混合研究范式，因此，本研究遵循行动研究、设计科学研究和 IIIE 的指导方针和原则。