Scholarlink

<p>In the presence of overdispersion, the spatial beta-binomial model can be a useful tool for modeling bounded count data. However, in practical applications, additional complexities such as unobserved heterogeneity, model misspecification, and non-standard distributional features may introduce extra non-spatial variation and shape characteristics that the beta-binomial model cannot fully accommodate. In this paper, we develop a flexible extension of this model that better captures distributional features while preserving an interpretable hierarchical structure. The proposed approach adapts the Beta-2-Binomial (B2B) formulation to spatial datasets by incorporating a shape parameter that captures extra distributional variability. The model accommodates spatial dependence via a latent Gaussian random field. Bayesian inference is performed using Markov chain Monte Carlo algorithms. Through simulation studies and an application to Loa loa prevalence data from Cameroon and Nigeria, we show that the proposed model provides improved predictive performance compared with existing spatial binomial approaches in the scenarios examined.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Flexible spatial regression for overdispersed bounded count data with distributional heterogeneity

  • 期刊论文

  • 发表日期：

    2026年2月20日

查看全文
  • Kosar Mahmood Hassan,
  • Majid Jafari Khaledi,
  • Esmaeil Najafi,
  • Atefeh Saboori

摘要

In the presence of overdispersion, the spatial beta-binomial model can be a useful tool for modeling bounded count data. However, in practical applications, additional complexities such as unobserved heterogeneity, model misspecification, and non-standard distributional features may introduce extra non-spatial variation and shape characteristics that the beta-binomial model cannot fully accommodate. In this paper, we develop a flexible extension of this model that better captures distributional features while preserving an interpretable hierarchical structure. The proposed approach adapts the Beta-2-Binomial (B2B) formulation to spatial datasets by incorporating a shape parameter that captures extra distributional variability. The model accommodates spatial dependence via a latent Gaussian random field. Bayesian inference is performed using Markov chain Monte Carlo algorithms. Through simulation studies and an application to Loa loa prevalence data from Cameroon and Nigeria, we show that the proposed model provides improved predictive performance compared with existing spatial binomial approaches in the scenarios examined.