Enhancing Healthcare With WBAN and Digital Twins: A Machine Learning Approach for Predictive Health Monitoring

The vital signs and medical condition telemetry from remote areas, such as the death rate, are urgent for hospital isolation. By recognizing and forecasting the causes of similar diseases, Wireless Body Area Networks (WBAN), made up of powerful wearable Sensor Nodes (SNs) located across the human body, generate a huge surge in data backed by other SNs. The collected data undergoes processing and is then sent to a remote medical server over the Internet. Machine learning (ML) has reinvented many paradigms, especially in healthcare, where it is a key resource in WBAN. A new advancement in this field is the integration of digital twins—virtual representations of the physical body. These digital twins can continuously simulate and predict health outcomes based on real-time WBAN data, enabling proactive healthcare interventions. This paper assessed several ML models based on their ability to process WBAN data, such as glucose, blood pressure, and body temperature, among others obtained from WBAN. These parameters feed into the digital twins, further refining the predictive and diagnostic capabilities of the models. The ML algorithms used include Logistic Regression (LR), Support Vector Classifier (SVC), K-Nearest Neighbours (KNN), Naive Bayes (NB), Decision Tree (DT), Random Forest (RF), Gradient Boosting (GB0, Neural Network (NN), AdaBoost (AB), Bagging (Ba), Extra Trees (ET), and XGBoost (XGB). Each model was evaluated based on accuracy, precision, recall, F1 score, ROC AUC score, and cross-validation accuracy. The RF model scored highest across all measurements, with accuracy (97%), precision (98%), recall (97%), F1 score (98%), ROC AUC score (97%), and cross-validation accuracy (97%). XGB and SVC also performed strongly, with GB, Ba, ET, and AB ranking high in efficiency, with consistent scores in the mid-90s. Integrating digital twins with these models offers an innovative framework to improve healthcare efficiency, allowing for real-time simulation and more accurate forecasting of patient outcomes.