Enhanced health monitoring in IoT with auto-metric graph neural networks and Archimedes optimisation

Abstract

ABSTRACT IoT-based healthcare monitoring systems often lack context-awareness, hindering their ability to provide personalised and accurate healthcare services. The proposed architecture addresses this challenge by incorporating auto-metric graph neural network with Archimedes optimisation algorithm is proposed in this paper for Health Care Monitoring in IoT-based Context-Aware Architecture (ANGNN-AOA-IoT-CA). The proposed work contains four phases: IoT phase, data preprocessing phase, context-aware phase and decision-making phase. In internet of things (IoT) phase, the sensor nodes are used for identifying the health status. Then, the amassed data are stored in storage layer. Hence, in the data preprocessing phase, the data redundancy, recurrence/repetition data are deleted. In context-aware phase, the preprocessed information is presented in the cloud and fog layer. As a result, the context-aware phase minimises the search space activities. In decision-making phase, the data are extracted and given to the ANGNN for classifying as normal and critical condition. Then Archimedes optimisation algorithm is utilised to acquire the better solution. The acquired outcomes of the proposed technique are analysed with existing models. Finally, the proposed method attains 2.37%, 2.95% and 1.17% higher accuracy, 1.09%, 1.47% and 1.53% higher sensitivity, 1.17%, 0.73% and 1.22% higher specificity compared with existing methods.