COVID-19 Pandemic Trend Prediction in America Using ARIMA Model

Abstract

COVID-19 trend prediction helps policymakers to handle disease situations. Therefore, it is necessary to predict the pandemic spread trend for prevention and control. The traditional infectious disease model is established according to the transmission characteristics of the disease. However, the trend prediction method of the traditional infectious disease model ignores considering the actual prevention and control situation, resulting in inaccurate models. To address this problem, this paper uses the ARIMA model to predict the spreading trend. First, we download the pandemic data from the website, compare the pandemic situation in different countries and select the United States as the research object. Second, the time series forecasting method is used to analyze the characteristics of the experimental data set. Finally, we use the ARIMA model to analyze the confirmed cases of COVID-19 in the United States and predict the spreading trend. To verify the effectiveness of the ARIMA model, we compare it with the prophet model and random forest model, evaluate the model performance with mean absolute scaled error, symmetric mean absolute percentage error, and root mean squared error. The experimental results illustrate that the ARIMA model significantly outperforms baselines by obtaining the three values of 0.14,9.97, 22316.57, respectively. The empirical results based on the pandemic spreading prediction in the United States show that the model has good applicability and accuracy.