Explainable death toll motion modeling: COVID-19 data-driven narratives.

Area of Science:

• Epidemiology
• Data Science
• Public Health

Background:

• The COVID-19 pandemic necessitated rapid policy decisions amidst significant uncertainty.
• Numerous predictive models were developed to guide public health actions.

Purpose of the Study:

• To develop intuitive, country-level COVID-19 motion models using machine learning.
• To utilize model explainability for data-driven insights into death toll dynamics.
• To inform policymakers and epidemiologists on outbreak drivers and intervention effectiveness.

Main Methods:

• Employed machine learning algorithms to create motion models.
• Modeled COVID-19 death tolls using velocity and acceleration metrics.
• Applied model explainability techniques to interpret model outputs.

Main Results:

• Developed country-level models describing COVID-19 death toll velocity and acceleration.
• Identified factors influencing the current pace of death tolls (velocity).
• Anticipated the impact of public health measures on future death toll deceleration (acceleration).

Conclusions:

• Machine learning-driven motion models offer valuable insights into COVID-19 dynamics.
• Model explainability enhances understanding of factors affecting death toll trends.
• These models support evidence-based policymaking and evaluation of public health strategies.