Prediction of COVID-19 Data Using Improved ARIMA-LSTM Hybrid Forecast Models

COVID-19 has developed into a global public health emergency and led to restrictions in numerous nations. Thousands of deaths have resulted from the infection millions individuals globally. Additionally, had significant impact on social economic activity around world. The elderly those with existing medical issues, however, are particularly vulnerable effects COVID-19. Pneumonia, acute respiratory distress syndrome, organ failure, death, etc. all possible outcomes severe cases... Traditional prediction approaches like ARIMA model multiple linear regression handle problem because new crown virus is process continual mutation. Deep learning models that can take account nonlinear elements include BP neural network LSTM prediction. To combine benefits traditional deep predictive create superior models, we blend models. When MSE, RMSE, MAE these three combined PSO-LSTM-ARIMA, MLR-LSTM-ARIMA, BPNN-LSTM-ARIMA, compared. We discovered third model, which included MAE, best accuracy. were selected for this investigation. begin, it employed single forecast pandemic data Germany. network, particle swarm method, then utilized merge it. corroborate finding, re-predicted epidemic Japan retrieved values BPNN-LSTM-ARIMA 6141895.956, 2478.285 1249.832. most accurate still integrated model. coupled offers highest effect, according our research. Combinatorial anticipate outbreak through study, aid governments authorities improving their responses educating about trends potential future directions. As result, industries enterprises may make better risk-management decisions protect safety operations personnel. It also helps healthcare facilities prepare deploy resources meet demands pandemic.