在這個研究中，我們建立一套貝氏時空廣義線性迴歸架構，結合 Potts 模型以捕捉交通事故發生率及其相關風險因子中的空間依存性與潛在的空間群集。推論方法採用 Metropolis-within-Gibbs 取樣演算法進行貝氏推斷。模擬研究結果顯示，本模型能有效辨識具有影響力的解釋變數，並能準確找出潛在的空間異質性結構及其對應之迴歸效果。

本研究進一步將模型應用於臺灣政府資料開放平台所提供之交通事故資料，涵蓋 2023 年 1 月至 2023 年 12 月的觀測期間。分析結果顯示，臺灣 349 個鄉鎮市區間的事故風險差異，主要源於關鍵影響因子之空間異質性。特別是，在臺灣中南部地區觀察到環境效應具有明顯的空間聚集現象。

In this research, we develop a Bayesian spatiotemporal generalized linear regression framework incorporating a Potts model to capture spatial dependence and latent regional clustering in traffic incidence and its associated risk factors. Bayesian inference is conducted via a Metropolis-within-Gibbs sampling algorithm. Results from simulation studies demonstrate the model's effectiveness in identifying influential covariates and in recovering latent spatial heterogeneity structures along with their corresponding regression effects.

The model is applied to traffic accident data from the Taiwan Open Government Data Platform, covering the period from January 2023 to December 2023. The analysis shows that variations in incident risk across 349 townships in Taiwan are attributable to spatial heterogeneity in key contributing factors. In particular, evidence of spatial clustering in environmental effects is observed in central and southern Taiwan.

S. Banerjee, B. P. Carlin, and A. E. Gelfand. Hierarchical modeling and analysis for spatial data. Chapman and Hall/CRC, 2003.
J. Besag. Spatial interaction and the statistical analysis of lattice systems. Journal of the Royal Statistical Society: Series B (Methodological), 36(2):192–225, 1974.
J. Besag, J. York, and A. Mollié. Bayesian image restoration, with two applications in spatial statistics. Annals of the institute of statistical mathematics, 43:1–20, 1991.
E. L. Frome and H. Checkoway. Use of poisson regression models in estimating incidence rates and ratios. American journal of epidemiology, 121(2):309–323, 1985.
A. Gelman and X.-L. Meng. Simulating normalizing constants: From importance sampling to bridge sampling to path sampling. Statistical science, pages 163–185, 1998.
A. Gelman, J. B. Carlin, H. S. Stern, and D. B. Rubin. Bayesian data analysis. Chapman and Hall/CRC, 1995.
E. I. George and R. E. McCulloch. Approaches for bayesian variable selection. Statistica sinica, pages 339–373, 1997.
M. Kim and L. Wang. Generalized spatially varying coefficient models. Journal of Computational and Graphical Statistics, 30(1):1–10, 2021.
B. G. Leroux, X. Lei, and N. Breslow. Estimation of disease rates in small areas: a new mixed model for spatial dependence. In Statistical models in epidemiology, the environment, and clinical trials, pages 179–191. Springer, 2000.
Y. Ma, M. Wang, and X. Chen. Identification and estimation of latent group structures in spatial autoregressive panels. Available at SSRN 4868881, 2024.
A. Mozdzen, A. Cremaschi, A. Cadonna, A. Guglielmi, and G. Kastner. Bayesian modeling and clustering for spatio-temporal areal data: an application to italian unemployment. Spatial Statistics, 52:100715, 2022.
D. Saha, P. Alluri, A. Gan, and W. Wu. Spatial analysis of macro-level bicycle crashes using the class of conditional autoregressive models. Accident Analysis & Prevention, 118:166–177, 2018.
C. Wang, M. A. Quddus, and S. G. Ison. Impact of traffic congestion on road accidents: A spatial analysis of the m25 motorway in england. Accident Analysis & Prevention, 41(4):798–808, 2009.
P. Xu and H. Huang. Modeling crash spatial heterogeneity: Random parameter versus geographically weighting. Accident Analysis & Prevention, 75:16–25, 2015.
P. Xu, H. Huang, N. Dong, and S. Wong. Revisiting crash spatial heterogeneity: a bayesian spatially varying coefficients approach. Accident Analysis & Prevention, 98:330–337, 2017.