Modeling temporal dependence in a sequence of spatial random partitions driven by spanning tree: an application to mosquito-borne diseases

2025-01-08

AI Summary

Identifying spatiotemporal patterns of mosquito-borne diseases faces the challenge of jointly respecting spatial structural constraints and temporal dependencies. Method: We propose a stochastic spanning tree–based spatiotemporal joint prior and develop a spatiotemporal Poisson mixture model with variable discrete parameters, enabling simultaneous Bayesian inference of dynamic spatial partitioning and disease intensity. The method integrates product partition models, stochastic spanning tree search, and MCMC sampling—overcoming limitations of conventional static partitioning or temporally independent modeling. Contribution/Results: Applied to weekly dengue fever data from southeastern Brazil (2018–2023), the model accurately captures the evolution of spatiotemporal clusters. Simulation studies demonstrate significantly higher clustering accuracy and parameter robustness compared to state-of-the-art methods, establishing a novel paradigm for dynamic spatial partitioning in infectious disease modeling.

Technology Category

• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Web Mining and Content Analysis: Models for Web evolution
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics

Abstract

Spatially constrained clustering is an important field of research, particularly when it involves changes over time. Partitioning a map is not simple since there is a vast number of possible partitions within the search space. In spatio-temporal clustering, this task becomes even more difficult, as we must consider sequences of partitions. Motivated by these challenges, we introduce a Bayesian model for time-dependent sequences of spatial random partitions by proposing a prior distribution based on product partition models that correlates partitions. Additionally, we employ random spanning trees to facilitate the exploration of the partition search space and to guarantee spatially constrained clustering. This work is motivated by a relevant applied problem: identifying spatial and temporal patterns of mosquito-borne diseases. Given the overdispersion present in this type of data, we introduce a spatio-temporal Poisson mixture model in which mean and dispersion parameters vary according to spatio-temporal covariates. The proposed model is applied to analyze the number of dengue cases reported weekly from 2018 to 2023 in the Southeast region of Brazil. We also evaluate model performance using simulated data. Overall, the proposed model has proven to be a competitive approach for analyzing the temporal evolution of spatial clustering.

Problem

• Model temporal dependence in spatial random partitions
• Identify spatio-temporal patterns of mosquito-borne diseases
• Analyze dengue cases using spatio-temporal Poisson mixture model

Innovation

• Bayesian model for time-dependent spatial partitions
• Random spanning trees for partition exploration
• Spatio-temporal Poisson mixture model for overdispersion