Dual-phase neural networks for feature extraction and ensemble learning for recognizing human health activities

Published: 01 Jan 2025, Last Modified: 05 Jun 2025Appl. Soft Comput. 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Human activity recognition aids elderly care, lifestyle monitoring, and healthcare.•Novel Dual-Phase Fused Network with ensemble learning for recognizing human activities.•DH-FusedCNN and Bi-GRU-Attention networks extract global-local HAR features.•Dual-Phase ensemble uses stacked weighted voting for robust HAR classification.•Experiments on six datasets achieve up to 99.66% accuracy across diverse tasks.
Loading