Count-valued time series models for COVID-19 daily death dynamics.

Area of Science:

• Biostatistics
• Epidemiology
• Time Series Analysis

Background:

• Daily COVID-19 death counts exhibit complex dynamics requiring advanced statistical modeling.
• Existing models may not fully capture the non-linear and time-varying nature of fatality data.

Purpose of the Study:

• To propose a generalized non-linear state-space model for analyzing count-valued time series of COVID-19 fatalities.
• To capture and model the dynamic changes in daily COVID-19 death counts.
• To validate and apply the proposed model to real-world COVID-19 data.

Main Methods:

• Development of a generalized non-linear state-space model with a latent state process incorporating second-order differencing and an AR(1)-ARCH(1) model.
• Fitting a series of Bayesian hierarchical models within the proposed framework.
• Model evaluation and comparison using predictive assessment on COVID-19 daily death counts from New York City boroughs.

Main Results:

• The proposed model elements were justified through rigorous model assessment.
• Bayesian hierarchical models within the framework effectively captured COVID-19 fatality dynamics.
• The model demonstrated applicability and effectiveness in analyzing COVID-19 death counts in New York City.

Conclusions:

• The generalized non-linear state-space model provides a robust framework for analyzing COVID-19 fatality time series.
• The model's components are validated, offering insights into epidemic dynamics.
• The framework successfully extends to analyze COVID-19 death counts in diverse geographical locations, such as Texas counties.