From COVID-19 to monkeypox: a novel predictive model for emerging infectious diseases

Area of Science:

• Epidemiology
• Public Health
• Computational Biology

Background:

• Emerging infectious diseases present major global health challenges.
• Accurate early forecasting is vital for resource allocation and emergency planning.
• Existing models often struggle in data-scarce scenarios.

Purpose of the Study:

• To develop a comprehensive predictive model for emerging infectious diseases.
• To enhance prediction accuracy and practicality using a novel approach.
• To improve forecasting capabilities in data-limited situations.

Main Methods:

• Integrated a blending framework, transfer learning, and incremental learning.
• Utilized the biological feature Rt (basic reproduction number).
• Transferred features from COVID-19 data to a monkeypox dataset, employing dynamic incremental learning.

Main Results:

• The blending framework excelled in short-term (7-day) predictions.
• Transfer and incremental learning improved Root Mean Square Error (RMSE) by 91.41% and Mean Absolute Error (MAE) by 89.13%.
• Inclusion of the Rt feature further refined RMSE by 1.91% and MAE by 2.17%.

Conclusions:

• Multimodel fusion and real-time data updates show significant potential for infectious disease prediction.
• The developed model offers enhanced adaptability and precision, particularly in data-scarce environments.
• This research provides theoretical insights and technical support for public health emergency responses.