Modeling unobserved heterogeneity using finite mixture random parameters for spatially correlated discrete count data

摘要: Abstract Road segments with identical site-specific attributes often exhibit significantly different crash counts due to unobserved reasons. The extent of heterogeneity associated a road feature is be estimated prior selecting the relevant safety treatment. Moreover, count data over-dispersed and spatially correlated. This paper proposes spatial negative binomial specification random parameters for modeling contiguous segments. incorporated using finite multi-variate normal mixture on parameters; this allows non-normality, skewness in distribution parameters, facilitates correlation across relaxes any distributional assumptions. model extracts inherent groups that are equally sensitive an average; within these also allowed proposed framework. simultaneously accounts potential from neighboring A Gibbs sampling framework leverages recent theoretical developments data-augmentation algorithms, elegantly sidesteps many computational difficulties usually Bayesian inference models. Empirical results suggests presence two latent study network. features effect were identified.