Semiparametric Bayesian inference on generalized linear measurement error models (Q1685290)

Generalized linear models (GLMs) are widely used to fit responses that do not satisfy the usual requirements of least-squares methods. GLMs with covariates having measurement errors (MEs) are often referred to as generalized linear measurement error models (GLMEMs). This paper introduces GLMEMs by using a centered DP mixture model to specify the distribution of covarate MEs. It develops a Bayesian MCMC algorithm to make Bayesian inference on GLMEMs by using the Gibbs sampler together with the Metropolis-Hastings algorithm. Two Bayesian case deletion diagnostic measures are presented to detect influential observations based on the K-L divergence and Cook's distance.