Hybrid Forecasting Model in Pandemic Prediction: Lesson from COVID-19

Predicting the spread of pandemic is crucial for public health planning and resource allocation. This work proposes a hybrid forecasting model (integrating polynomial and multivariate linear regression) to forecast the spread of COVID-19 pandemic. The model achieves an R 2 value of 0.98, outperforming ARIMA (0.97), multivariate linear regression (0.91), and polynomial regression (0.21). The variables like daily test cases, vaccinations, and new cases enhance the accuracy of prediction. It is effective for COVID-19. However, for other pandemics with different characteristics of transmission or other variables such as changes in policy and mutations of the virus, the model may not predict accurately, but it still encourages to tackle those. The developed model is robust in terms of short-term to medium-term forecast.