Long Short-Term Memory Prediction for COVID19 Time Series

Entire world has been dealing with the number of new Coronavirus 2 or COVID-19 cases. The spread of thissevere acute respiratory syndrome has produced manyconcerns worldwide. Having data related to coronavirusavailable for tests, novel models for forecasting the number ofnew cases can be developed. In this paper, a long short-termmemory (LSTM) based methodology is applied for suchprediction. Here, experimental analysis is performed with theparameters, such as the number of layers and units of thenetwork. The root mean squared error is calculated for datacorresponding to the Republic of Serbia, as well as perdifferent continents. The results show that LSTM model canbe useful for further analysis and time series prediction © 2021, Telfor Journal. All Rights Reserved.

Long Short-Term Memory Forecasting for COVID19 Data

Nowadays everyone is talking about Coronavirus 2 or COVID-19. It is a severe acute respiratory syndrome which produces a lot of concerns around the globe. Since data is available for everyone, as well as for the Republic of Serbia, we used it for experiments via a neural network. Here, a long short-term memory approach is applied in order to make experiments with its parameters, such as the number of layers and the number of hidden units. The results show proper modelling from the standard root mean square error standpoint. The paper contribution is related to testing the parameters in the long short-term memory approach for recent data from the Republic of Serbia.