Pelican Gorilla Troop Optimization Based on Deep Feed Forward Neural Network for Human Activity Abnormality Detection in Smart Spaces

Abstract

Healthcare management can be improved using artificial intelligence-powered Internet of Things (IoT) for remotely collecting and analyzing medical data. Home-based IoT healthcare proved its effectiveness in helping the elderly and people with special care needs enjoy safer and more independent living. Deep learning techniques improve the management of healthcare systems through intelligent analysis and real-time tracking of health indicators and auto-administering medication. This paper proposes a novel pelican gorilla troop optimization-assisted deep feed-forward neural network for health indicators abnormality detection. As deep learning requires a significant data dimension to provide reliable results, the data augmentation process is carried out through the bootstrapping approach to improve abnormality detection. Furthermore, Z-score normalization is used to complete data pre-processing and achieve better detection outcomes. The experimental evaluation results show that the proposed solution realizes a detection performance in terms of accuracy, precision, and recall achieving of 0.879, 0.902, and 0.929, respectively.