Predicting Daily Confirmed Cases in Midwestern Central States in U.S. by Using AIMA and LSTM

Covid-19 is an epidemic disease caused by SARS-Cov-2 virus, which is a type of coronavirus. This virus is highly contiguous, and the confirmed cases of this disease have increased rapidly in a short period. After one month of the first reported case, the World Health Organization (WHO) claims that the Covid-19 will become an international public health emergency. The main purpose of this thesis is to predict the daily confirmed cases of Covid-19 in the midwestern central states in the U.S, by using Autoregression Integrated Moving Average (ARIMA) model and Long Short-Term Memory network (LSTM), which is a type of recurrent neural network. We compare the Root Mean Square Error (RMSE) for the prediction to determine the performance of the two methods. In this thesis, we show that the LSTM network has a smaller prediction RMSE. Also, both models capture the seasonality of the dataset. LSTM captures the trend of the dataset and has a higher prediction than expected values. ARIMA does not capture the trend of the dataset and will have a larger range. Therefore, we can conclude that LSTM is a better method for predicting daily confirmed cases of Covid-19 in the Midwestern central states in the U.S