Grey forecasting models based on internal optimization for Novel Corona virus (COVID-19)

Area of Science:

• Epidemiology
• Data Science
• Mathematical Modeling

Background:

• Accurate pandemic forecasting is challenged by limited data.
• The Novel Coronavirus pandemic necessitates reliable forecasting tools.
• Grey forecasting models offer a potential solution for data-scarce scenarios.

Purpose of the Study:

• To propose two novel Internally Optimized Grey Prediction Models (IOGMs).
• To enhance the conventional Grey Forecasting model (GM(1,1)) for pandemic spread prediction.
• To evaluate the performance of IOGMs in forecasting COVID-19 cases across various Indian states.

Main Methods:

• Modification of the conventional Grey Forecasting model (GM(1,1)).
• Development of two IOGMs by stacking infected case data with diverse overlap periods.
• Comparative analysis against conventional GM(1,1) and NGM(1,1,k) models using time series data from Indian states.

Main Results:

• The proposed IOGMs demonstrated satisfactory prediction accuracies for pandemic spread.
• Forecasted results from IOGMs aligned well with mean infected case data.
• Error index evaluations indicated that models with higher overlap periods achieved superior performance.

Conclusions:

• Internally Optimized Grey Prediction Models offer a viable approach for pandemic forecasting, especially with limited data.
• The choice of overlap period significantly impacts the accuracy of grey prediction models.
• IOGMs provide valuable insights for public health authorities managing infectious disease outbreaks.