AutoML-Driven Intelligent Big Data Analytics for Scalable and Adaptive Enterprise Decision-Making Systems

Abstract

The fast growth of enterprise-level data has created an urgent demand for scalable and flexible analytics solutions that can support real-time decision-making. Traditional data mining workflows' high reliance on human interaction for tasks like data preparation, features optimization, models selection, and hyper-parameter tuning limits their efficacy and adaptability in rapidly evolving organizational contexts. To address these concerns, this article proposes an AutoML-driven intelligent analytics system that automated the whole data mining pipeline within corporate big data environments. The proposed architecture incorporates automated processing, feature selection, model development, and hyper-parameter optimization inside distributed computing infrastructure. A meta-learning module allows for adaptive choosing models based on the inherent characteristics of the dataset, and an integrated optimization technique ensures a balance between predicted accuracy and computing cost. Analytical analyses and architectural representations that highlight the connections between efficiency, capacity, and adaptability are used to assess the framework. The results demonstrate how well hybrid methods that combine automated model creation, drift detection, and meta-learning work to create reliable and self-optimizing machine learning algorithms. This paper advances adaptive intelligence by highlighting the relevance of dynamic architectures in forming next-generation enterprise analytics and by identifying current research gaps and suggesting future paths.