A partially linear additive model for clustered proportion data

• DOI: 10.1002/sim.7573
• Publication Date: 2017-12-16T21:27:27Z
• Authors: Weihua Zhao; Heng Lian; Dipankar Bandyopa...
• Keywords: Spline (mechanical); Additive model; Generalized additive model; Estimating equations

Proportion data with support lying in the interval [0,1] are a commonplace various domains of medicine and public health. When these available as clusters, it is important to correctly incorporate within‐cluster correlation improve estimation efficiency while conducting regression‐based risk evaluation. Furthermore, covariates may exhibit nonlinear relationship (proportion) responses quantifying disease status. As an alternative existing classical methods for modeling proportion (such augmented Beta regression) that uses maximum likelihood, or generalized estimating equations, we develop partially linear additive model based on quadratic inference function. Relying quasi‐likelihood techniques polynomial spline approximation unknown nonparametric functions, obtain estimators both parametric part our study their large‐sample theoretical properties. We illustrate advantages usefulness proposition over other alternatives via extensive simulation studies, application real dataset from clinical periodontal study.