A0881
Title: A functional semiparametric mixed effects state space model with prior information for county level spatiotemporal data
Authors: Mengying You - Shanghai University of International Business and Economics (China) [presenting]
Abstract: A functional semiparametric mixed-effects state space model is proposed for analyzing COVID-19 case dynamics at the U.S. county level, incorporating both spatial structure and prior information. The model captures temporal trends using general smoothing splines and introduces spatially structured random effects to borrow strength across geographically or epidemiologically similar counties. A state-space formulation enables efficient recursive estimation and forecasting while accommodating both latent temporal processes and county-specific variability. The inclusion of prior knowledge on curve shapes enhances model stability in regions with sparse or noisy data. Through comparison with standard spline-based time series models, the approach demonstrates improved predictive accuracy, reduced forecast uncertainty, and greater capacity to recover shared epidemic patterns across counties. This framework is broadly applicable to spatiotemporal epidemiological surveillance and regional policy evaluation.
