A Methodological Approach for Predicting COVID-19 Epidemic Using EEMD-ANN Hybrid Model

Area of Science:

• Epidemiology
• Data Science
• Computational Biology

Background:

• The COVID-19 pandemic, caused by the novel coronavirus, presents significant challenges for global healthcare systems.
• Accurate prediction of epidemic spread is crucial for effective resource allocation and public health interventions.
• Limited data and complex dynamics make COVID-19 forecasting a difficult task.

Purpose of the Study:

• To develop and evaluate a hybrid model for predicting the COVID-19 epidemic.
• To enhance prediction accuracy by integrating Ensemble Empirical Mode Decomposition (EEMD) with Artificial Neural Networks (ANN).
• To provide a tool for healthcare management and preventive action.

Main Methods:

• A hybrid model combining EEMD and ANN was proposed.
• Real-time COVID-19 time-series data from January 22, 2020, to May 18, 2020, was utilized.
• EEMD was employed for data denoising and decomposition into sub-signals, followed by ANN training.

Main Results:

• The proposed hybrid EEMD-ANN model demonstrated superior performance compared to traditional statistical analysis methods.
• The model effectively predicted COVID-19 epidemic trends using denoised time-series data.
• The results indicate the model's potential for reliable epidemic forecasting.

Conclusions:

• The hybrid EEMD-ANN model shows significant promise for accurate COVID-19 epidemic prediction.
• This predictive capability can assist governments and healthcare providers in planning and implementing timely interventions.
• The study highlights the value of advanced computational models in managing public health crises.