Leveraging Machine Learning for Personalized Wearable Biomedical Devices: A Review

Abstract

This review investigates the convergence of artificial intelligence (AI) and personalized health monitoring through wearable devices, classifying them into three distinct categories: bio-electrical, bio-impedance and electro-chemical, and electro-mechanical. Wearable devices have emerged as promising tools for personalized health monitoring, utilizing machine learning to distill meaningful insights from the expansive datasets they capture. Within the bio-electrical category, these devices employ biosignal data, such as electrocardiograms (ECGs), electromyograms (EMGs), electroencephalograms (EEGs), etc., to monitor and assess health. The bio-impedance and electro-chemical category focuses on devices measuring physiological signals, including glucose levels and electrolytes, offering a holistic understanding of the wearer’s physiological state. Lastly, the electro-mechanical category encompasses devices designed to capture motion and physical activity data, providing valuable insights into an individual’s physical activity and behavior. This review critically evaluates the integration of machine learning algorithms within these wearable devices, illuminating their potential to revolutionize healthcare. Emphasizing early detection, timely intervention, and the provision of personalized lifestyle recommendations, the paper outlines how the amalgamation of advanced machine learning techniques with wearable devices can pave the way for more effective and individualized healthcare solutions. The exploration of this intersection promises a paradigm shift, heralding a new era in healthcare innovation and personalized well-being.

Development of a Personalized Multiclass Classification Model to Detect Blood Pressure Variations Associated with Physical or Cognitive Workload

“…Although DL and ML have found extensive application in BP assessment, the considerable inter-subject variability has posed challenges in formulating a sufficiently generalized model whose performance could also be maintained outside of the initial dataset. Therefore, drawing inspiration from established practices in the field of human activity recognition [ 34 , 35 ], numerous studies have suggested the formulation of person-specific models for the examination of this clinical parameter [ 36 , 37 , 38 , 39 ].…”