Page 116 - MI-2-3

Microbes & Immunity

ORIGINAL RESEARCH ARTICLE
A comprehensive statistical analysis of COVID-19

trends: Global and United States insights
through ARIMA, regression, and spatial models

Zhihao Lei *
1,2
1 School of Mathematics, University of Edinburgh, Edinburgh, Scotland, United Kingdom
2 Department of Biostatistics, School of Public Health, Brown University, Providence, Rhode Island,
United States of America

Abstract

The COVID-19 pandemic has driven the need for accurate data analysis and
forecasting to support public health decision-making. This study applied
autoregressive integrated moving average (ARIMA) models and ARIMA models
with exogenous variables to predict short-term trends in confirmed COVID-19 cases
across several regions, including the United States of America, Asia, Europe, and
Africa. Model performance was compared between ARIMA and the automated
model selection function, auto.arima, and anomaly detection was performed
to investigate discrepancies between predicted and observed case numbers.
Additionally, the study explored the relationship between vaccination rates and
*Corresponding author: new case trends while also examining the influence of socioeconomic factors—such
Zhihao Lei as gross domestic product per capita, human development index, and healthcare
(Z.Lei-6@sms.ed.ac.uk) resources availability—on COVID-19 incidence across countries. The findings
Citation: Lei Z. A comprehensive provide valuable insights into the effectiveness of predictive models and highlight
statistical analysis of COVID-19 the significant role of socioeconomic factors in the spread of the virus, thereby
trends: Global and United States contributing to the development of more effective strategies for future epidemic
insights through ARIMA,
regression, and spatial models. prevention and control.
Microbes & Immunity.
2025;2(3):108-129.
doi: 10.36922/MI025040007 Keywords: Autoregressive integrated moving average model; COVID-19; Public health;
Received: January 22, 2025 Socioeconomic factors; Time series forecasting; Vaccination rates
Revised: April 9, 2025
Accepted: May 12, 2025 1. Introduction
Published online: June 18, 2025
Since the onset of the COVID-19 pandemic in late 2019, the pandemic has had profound
Copyright: © 2025 Author(s).
This is an Open-Access article and widespread effects on global public health, economies, and daily life. As of 2024, it
distributed under the terms of the continues to pose challenges to healthcare systems worldwide, underscoring the ongoing
Creative Commons Attribution
License, permitting distribution, need for accurate forecasting of case trends for effective policy-making decisions and
and reproduction in any medium, intervention strategies. Statistical modeling, particularly time series analysis, has proven
provided the original work is
properly cited. to be a valuable tool in predicting the trajectory of the pandemic and supporting the
development of effective public health responses. 1
Publisher’s Note: AccScience
Publishing remains neutral with Among the range of statistical models, the autoregressive integrated moving average
regard to jurisdictional claims in
published maps and institutional (ARIMA) model has been widely employed in epidemiological studies for short-term
2
affiliations. forecasting due to its simplicity and effectiveness in modeling temporal data. ARIMA

Volume 2 Issue 3 (2025) 108 doi: 10.36922/MI025040007