School of Mathematical & Statistical Sciences Faculty Publications
Document Type
Article
Publication Date
7-22-2026
Abstract
The COVID-19 pandemic highlighted the need for accurate epidemic forecasting to support public health decision-making. Most existing approaches depend heavily on human mobility data, while largely neglecting population behavior shaped by socio-cultural norms. In this study, we analyze daily COVID-19 mortality and Google mobility data from 72 countries during the first 130 d of the pandemic, a period characterized by high uncertainty and behavioral heterogeneity. In particular, we examine whether Hofstede’s country-level cultural dimensions can serve as latent behavioral forecasters of mortality in lieu of dynamic mobility indicators. Using 100 d for training and 30 d for forecasting, we employ three deep learning frameworks: an autoregressive model of mortality dynamics, a sequence-to-sequence model incorporating mobility patterns, and a model that directly integrates static cultural features into mortality trajectories. Our results show that while mobility data generally improves long-term forecasting, models based solely on cultural dimensions achieve comparable and in some cases superior predictive performance. Among the six cultural dimensions, individualism and uncertainty avoidance emerge as the strongest protective predictors of mortality trends. Although no causal inference is intended, these findings demonstrate that fixed socio-cultural factors capture population-level behavioral information that is not substantially different from that provided by time-varying mobility data, suggesting a viable and informative alternative for epidemic forecasting when mobility data are unavailable or costly.
Recommended Citation
Abbas, Saif, Tamer Oraby, Michael G. Tyshenko, and Samit Bhattacharyya. “Cultural Traits May Replace Human Mobility Data in Forecasting COVID-19 Mortality: A Deep Learning Approach.” Machine Learning: Health 2, no. 2 (2026): 025008. https://doi.org/10.1088/3049-477X/ae87a9.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.
DOI
10.1088/3049-477X/ae87a9
Comments
© 2026 The Author(s). Published by IOP Publishing Ltd.
Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 license. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.