Reduced Spread Risk of COVID-19 based on Tracking System Monitoring in WSNs

The phrase "COVID-19" has become the most popular and highly searched term on Google since its emergence in November 2019. The world must combat this pandemic. With advancements in mobile technology and sensors, Developing Internet of Things (IoT)-based healthcare solutions is now feasible. Proactive as well as preventative, these new systems, are in contrast to the traditional reactive healthcare systems. In this study, a paradigm based on the Internet of Things is proposed for detecting COVID-19 suspects. The framework presented in this research focuses on predicting COVID-19 in its early stages. To determine whether the virus is present, it uses health measurements collected from sensors and other Internet of Things devices. By gathering symptom data related to COVID-19, the behavior of the virus can be better understood. Four machine learning models are introduced in this paper: CatBoost, Decision Tree, Random Forest, and Logistic Regression. These models are used to create a Diagnostic Learning Classifier (DLC) model. The top three performing models, Decision Tree, Logistic Regression, and CatBoost, are combined using a weighted average or majority voting approach, depending on the nature of the problem. The DLC model is compared to base models such as Random Forest, Decision Tree, Logistic Regression, and CatBoost, using multiple evaluation criteria. The goal of this ensemble model is to increase forecast accuracy by utilizing the variety and advantages of the chosen models. It delivers great diagnostic accuracy when used as a machine learning model on the gathered data. Notably, the DLC classifier achieves an accuracy of 91%, followed by Random Forest (86%), CatBoost (85%), Logistic Regression (84%), and Decision Tree (82%). These results demonstrate the effectiveness of the voting classifier in efficiently managing COVID-19 by accurately diagnosing patients and notifying health authorities of suspected cases.