Predictive and analysis of COVID-19 cases cumulative total：ARIMA model based on machine learning

ABSTRACT

At present, COVID-19 poses a serious threat to global human health, and the cumulative confirmed cases in America, Brazil and India continue to grow rapidly. Therefore, the prediction models of cumulative confirmed cases in America, Brazil and India from August 1, 2021 to December 31, 2021 were established. In this study, the prevalence data of COVID-19 from 1 August 2021 to 31 December 2021 were collected from the World Health Organization website. Several ARIMA models were formulated with different ARIMA parameters. ARIMA (7,2,0), ARIMA (3,2,1), and ARIMA (10,2,4) models were selected as the best models for America, Brazil, and India, respectively. Initial combinations of model parameters were selected using the automated ARIMA model, and the optimized model parameters were then found based on Bayesian information criterion (BIC). The analytical tools autocorrelation function (ACF), and partial autocorrelation function (PACF) were used to evaluate the reliability of the model. The performance of different models in predicting confirmed cases from January 1, 2022 to January 5, 2022 was compared by using root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE). This study shows that ARIMA models are suitable for predicting the prevalence of COVID-19 in the future. The results of the analysis can shed light on understanding the trends of the outbreak and give an idea of the epidemiological stage of these regions. Besides, the prediction of COVID-19 prevalence trends of America, Brazil, and India can help take precautions and policy formulation for this epidemic in other countries.