“…The training phase of the model collects data variables primarily based on publicly available health conditions datasets. These variables include data about blood pressure, pulse oximetry, the concentration of glucose in the blood, activity tracking, sleep tracking with the corresponding prediction class; in addition to patient feedback and evaluation data (Dinsmore et al, 2021 ); finally, data related to health trajectories from Healthcare Administrative Databases (HADs) such as diagnoses and medication prescriptions are also included (Veronica et al, 2022 ). The production phase of the model collects data coming from the proposed tool and its associated sensory fabrics, as described in Figures 2 , 4 to predict emergent health situations requiring interventions such as early hospital admission, diagnosis, clinical procedures, and medications, will be compared with deep learning (DL) and convolutional neural network (CNN) (Nguyen et al, 2016 ; Pham et al, 2016 ) vs. traditional machine learning algorithms, e.g., Bayesian probabilistic model, k nearest neighbors, logistic regression, support vector machines, and decision tree (Deparis et al, 2018 ; Hansen et al, 2018 ; Khalid et al, 2018 ; Noh et al, 2019 ; Ben-Assuli and Padman, 2020 ; Franz et al, 2020 ; Veronica et al, 2022 ).…”