Integration of Wireless Network Features with Human Activity Recognition: A Comprehensive Dataset Analysis

Human activity recognition is a modern research field that aims to monitor human vitals throughout various activities. This field, however, is still under development in terms of its integration with other fields. This work combines wireless network data with human activity recognition (HAR) through the UCI-HAR dataset, which extends human activity recognition with wireless sensor networks(WSN) data from the University of California, Irvine (UCI). UCI-HAR is a machine learning repository that collects wireless sensor network data, and HAR data from 30 subjects is used throughout daily activities. This work bridges HAR with wireless network quality measurement. We conducted our experiments using Google Colab for six classes of the dataset. The results show that random forest had an accuracy of 97% across all six classes of the dataset, and 1.00 precision scores were achieved for classes 0, 3, 4, and 5, and low for class 6. The findings of this study are significant for the Internet of Things (IoT) applications in healthcare monitoring and smart home environments. It will contribute to developing intelligent and contextual-aware systems; therefore, it can improve various applications.