Event Specific Transmission Prognosis Gleaned from Machine Learning for COVID-19 Prophylaxis in Air-Conditioned Office

EVENT SPECIFIC TRANSMISSION PROGNOSIS GLEANED FROM MACHINE LEARNING FOR COVID-19 PROPHYLAXIS IN AIR-CONDITION OFFICE

Abstract

The unprecedented rate of emerging and re-emerging cases of COVID-19 with novel variants and the recent coining of long COVID-19 makes it essential to explore the transmission of SARS-CoV-2. An air-conditioned office room has been investigated for understanding the event-specific possibility of viral transmission of Coronavirus in composite climatic conditions. This work contains newly developed machine learning models to forecast the R-Event using artificial neural network (ANN), Gaussian Process Regression (GPR), Support Vector Machine (SVM), and Curve-Fitting (CF). Real-time monitoring of an air-conditioned office room was executed in the months of April and May of 2022 to collect the data. The proposed models were evaluated for their performances. For determining the merit of the four models, statistical parameters, namely the correlation coefficient, mean absolute error, root mean square error, mean absolute percentage error, Nash-Sutcliffe efficiency index, and a20-index, were considered. Eight input features were used to predict the target R-Event in this study. The main objective of this study is to develop a model to link CO₂ concentration with R-Event value for forecasting event specific infections in air-conditioned office environment. Results indicate that the developed ANN prediction model is the best among the four models to forecast the R-Event.