Forecasting emergency medicine reserve demand with a novel decomposition-ensemble methodology

Area of Science:

• Emergency Medicine
• Data Science
• Forecasting

Background:

• Effective emergency medicine reserve management is crucial, especially during public health events.
• Traditional demand forecasting methods struggle with complex, multi-component data influenced by various factors.

Purpose of the Study:

• To propose a novel forecasting model for emergency medicine reserve management.
• To address the limitations of traditional methods in handling complex, non-linear time-series data.

Main Methods:

• Empirical Mode Decomposition (EMD) to break down time-series data into intrinsic mode functions (IMFs).
• Elman neural networks and ARIMA models to forecast individual IMFs.
• Integration of component forecasts to generate final predictions.

Main Results:

• The proposed EMD-ELMAN-ARIMA (ELA) model demonstrated superior predictive accuracy.
• Empirical validation using Beijing influenza data (2014-2018) confirmed ELA's effectiveness.
• ELA outperformed standalone ARIMA and Elman models in forecasting accuracy.

Conclusions:

• The ELA model offers a robust solution for emergency medicine reserve demand forecasting.
• Decomposition of complex data series enhances prediction accuracy.
• This approach is vital for optimizing resource allocation during public health crises.