DPGLM: A Semiparametric Bayesian GLM with Inhomogeneous Normalized Random Measures

TL;DR

This paper presents DPGLM, a novel semiparametric Bayesian generalized linear model that incorporates an inhomogeneous normalized random measure, enabling flexible modeling of baseline distributions with validated simulation and real data applications.

Contribution

It introduces a new inhomogeneous normalized random measure-based prior for semiparametric Bayesian GLMs, extending existing models with a nonparametric baseline distribution.

Findings

Effective posterior simulation methods developed.

Model validated through extensive simulations.

Applied successfully to speech intelligibility data.

Abstract

We introduce a novel varying-weight dependent Dirichlet process (DDP) model that extends a recently developed semi-parametric generalized linear model (SPGLM) by adding a nonparametric Bayesian prior on the baseline distribution of the GLM. We show that the resulting model takes the form of an inhomogeneous completely random measure that arises from exponential tilting of a normalized completely random measure. Building on familiar posterior sampling methods for mixtures with respect to normalized random measures, we introduce posterior simulation in the resulting model. We validate the proposed methodology through extensive simulation studies and illustrate its application using data from a speech intelligibility study.