Predicting the infection level of COVID-19 virus using normal distribution-based approximation model and PSO

The infectious disease brought on by COVID-19 is the pandemic that has put the lives and economy of the entire planet in jeopardy. The only alternatives available before the creation of a vaccine are how to take preventive measures and how to best employ the restricted medical equipment following patient need. Due to the lead durations in vaccine development, this will hold for potential pandemics in the future. An accurate estimation of the Infection Level (IL) in a particular patient is essential to determine if a patient needs to be in quarantine, isolation, or a ventilator. In the case of COVID-19, however, physical contact between a patient and a doctor or any other person is extremely dangerous; as a result, an accurate prediction model is essential for estimating the IL without any physical contact. In this study, innovative mathematical model for estimating the intensity of infection in the human body after a specified number of days since the first symptom appeared is proposed. Other factors taken into account by this model include age, gender, chronic illnesses, smoking, and Body Mass Index (BMI). The model is made to be flexible so that it may accept data input for any country and produce results based on the infection pattern of that nation. Additionally, by utilising the normal distribution, it may cumulatively display the outcome for the entire world. The predicted error’s Root Mean Square Error (RMSE) value falls between 0.9% and 2.4%. In addition, the authors suggested an evolutionary strategy called Particle Swarm Optimisation (PSO) on the dataset used to optimise the variables in the mathematical technique that depicts how an IL spreads through a community, and then they compared the outcomes with the mathematical model. Although both approaches are effective, the evolutionary strategy outperforms the mathematical model.