Prediction of Covid19 using Neural Networks and Compartmental Models in Epidemiology

Abstract

The buzz word in the whole health sector in the world was COVID-19 alias CORONA virus. Present all over the world most rapid spread virus from one person to another person is the corona. It was initially started at Wuhan in China in December 2019; exponentially spread it to the entire country of China. Past three months, the virus was transformed as a significant outbreak in the world. Our study focuses on India covid19 infected cases. Sickness classification is one of the significant and more time taking task in medical diagnosis system.

To forecast the epidemic peak of coronavirus, which might help us to act appropriately to reduce the epidemic risk? Multilayer perceptron (MLP) with the backpropagation technique used on the data set to predict the covid19. A threshold value is controlled by ANN to distinguish positive and negative cases, and the model groups the experiments, either positive or negative, given the threshold. Compartmental model in epidemiology SEIR method to forecast the number of confirmed cases based on the current scenario and population. Neural network multilayer perceptron method with Backpropagation algorithm predicts the confirmation cased based on symptoms of the patient with an accuracy of 76%. SEIR model forecast the confirmed cases of India in the come days and starts decrease of infected cases end of the year 2020. Our study can predict the infected patient based on symptoms using neural network MLP model, and We are estimating the number of infected people in the coming days using the SEIR method.