Integration of Machine Learning and Big Data for Intelligent Cloud-Based Decision Support

Abstract

Organizations have invested heavily in big data platforms and machine learning (ML), yet many ML models never reach production, and many that do fail to improve decisions. The problem is rarely the algorithm alone. It lies in how data, models, and decisions are connected over time: data pipelines that silently change, models that drift, predictions that stop short of recommending action, and outputs that never enter the processes where decisions are made. This article proposes a closed-loop framework for integrating ML and big data into intelligent cloud-based decision support. The framework covers a hybrid cloud data foundation for sensitive data, validated data and feature pipelines, managed model development and serving, a decision layer that moves from prediction to prescription, a retrieval-augmented natural-language decision assistant, workflow-driven decision execution, and monitoring and autonomous operations that feed outcomes back into learning. The article also defines four levels of decision support, from descriptive to prescriptive, and a production-readiness checklist for decision models. Using a design-oriented approach grounded in ten studies published between 2017 and 2025, it illustrates the framework with hospital capacity planning. It argues that intelligent decision support depends on treating data, models, and decisions as one continuously operated loop rather than as separate projects.