Sensitivity analysis of disease-information coupling propagation dynamics model parameters.

Area of Science:

• Epidemiology
• Network Science
• Computational Social Science

Background:

• The disease-information coupling propagation dynamics model is vital for studying infectious disease spread.
• Parameter settings and sensitivity analysis are often overlooked, leading to significant errors in model results.
• Identifying key model parameters is essential for understanding coupling mechanisms and effective disease control.

Purpose of the Study:

• To explore influencing factors and identify key parameters in the disease-information coupling propagation dynamics model.
• To conduct global sensitivity analysis on six input parameters under both same and heterogeneous interaction radii.
• To enhance understanding of the model's coupling mechanism and inform disease control recommendations.

Main Methods:

• Sobol global sensitivity analysis algorithm was employed.
• Analysis was performed on six input parameters: cross-regional jump probability, information dissemination rate (βI), information recovery rate (μI), epidemic transmission rate (βE), epidemic recovery rate (μE), and preventive action probability.
• Simulations were conducted for both same and heterogeneous interaction radius scenarios.

Main Results:

• In the same interaction radius model, information dissemination rate (βI) and recovery rate (μI) most significantly impacted information spread and peak AI node density.
• In the heterogeneous interaction radius model, information spread rate (βI), disease recovery rate (μE), and information recovery rate (μI) were most influential on information layer dynamics.
• Disease transmission rate (βE), disease recovery rate (μE), and cross-regional jump probability (pjump) significantly affected disease layer dynamics (SE node density and transmission scale) in both interaction radius scenarios.

Conclusions:

• Sensitivity analysis reveals critical parameters governing disease and information spread in coupled models.
• Different interaction radii alter the influence of parameters on disease and information dynamics.
• Accurate parameter identification is crucial for reliable disease modeling and effective public health interventions.