Forecasting of Covid-19 Using Roll Prediction Based on Deep Learning

Corona Virus Disease 2019 (COVID-19), short for “new coronavirus pneumonia” and named “2019 coronavirus disease” by the World Health organization, refers to pneumonia caused by the new 2019 coronavirus infection. COVID-19 is highly transmissible and is spreading rapidly around the world, posing a threat to human health and already having a huge impact on the world economy. Therefore, it is very important and necessary to predict the trend of the number of patients with COVID-19. The present work proposes learning models. There have been many previous studies using deep learning neural networks to predict time series, and the results have been shown to be outstanding. The present work will change the forecast format from using the entire segment of the data to forecast a day or segment to a roller forecast format. We will start from Convolutional Neural Network (CNN) and then use simple Recurrent Neural Networks (RNN) and The Gated Recurrent Units (GRUs) cells along with Long Short-Term Memory (LSTM) cells to predict the deaths and cases in US. We have used publicly available data from John Hopkins University’s COVID-19 database. In this study we will find the advantages and disadvantages of roller prediction and the limitations of different models and predict future trends, which will play a positive role in the research and prediction of the prevention and control of the epidemic.