Design and Evaluation of a Multi-Metric Machine Learning Model for Human Activities Reorganization with Emphasis on Specificity and Real Time Adaptability

Abstract

Human activity recognition (HAR) has emerged as a crucial area of research due to its widespread applications in various domains, including healthcare, smart environments, and assistive technologies. With the proliferation of wearable sensors and the Internet of Things (IoT), the ability to accurately sense and interpret human activities has become increasingly important. Machine learning models have played a pivotal role in advancing HAR systems, enabling the effective recognition of complex activities from sensor data. This research paper provides a comprehensive review of machine learning models employed for human activity recognition, encompassing both traditional techniques and state-of-the-art deep learning approaches as well as a proposed approach. It discusses the challenges and considerations involved in activity recognition, such as data acquisition, feature extraction, and model selection. Additionally, the paper presents a comparative analysis of various machine learning models, evaluating their performance, strengths, and limitations across different activity recognition tasks and datasets and comparison among different machine learning models and proposed model.