Digital Ecosystem Architecture For Distributed Intelligence And Integrated Data Management In Smart Healthcare Systems

Abstract

Digital ecosystems are contributing to the evolution of smart health care through distributed intelligence and integration of data from interconnected medical ecosystems. However, data heterogeneity, high dimensionality, and inadequate integration pose major problems that hinder decision making at an appropriate level. In this research work, we propose an architecture that facilitate data acquisition, processing, and analytics. The dataset consists of 4000 records with 18 attributes collected from Internet of Things (IoT)-enabled healthcare environments, including physiological signals, wearable device data, and electronic health indicators, with a target variable representing patient health status. Min–max normalization is applied for uniform feature scaling, and Principal Component Analysis (PCA) is used to reduce dimensionality while preserving essential variance. An Intelligent Marine Predators Optimized Stacked Long Short-Term Memory (IntMP-SLSTM) model is developed and implemented using Python. The IntMP optimization enhances hyperparameter tuning and convergence, while the stacked LSTM captures long-term dependencies in sequential healthcare data. Experimental results achieved 98.92% accuracy, 98.94% precision, 98.89% recall, 98.93% specificity, and 98.94% F1-score, confirming its effectiveness for advanced smart healthcare applications. The proposed IntMP-SLSTM model has effectively addressed dimensionality and data heterogeneity, resulting in extremely accurate predictions in the field of smart health care.