Varying coefficient models as Mixed Models : reparametrization methods and bayesian estimation

Freni Sterrantino, Anna (2013) Varying coefficient models as Mixed Models : reparametrization methods and bayesian estimation. Bologna, IT: Dipartimento di Scienze Statistiche "Paolo Fortunati", Alma Mater Studiorum Università di Bologna, p. 18. DOI 10.6092/unibo/amsacta/3870. In: Quaderni di Dipartimento. Serie Ricerche (5). ISSN 1973-9346.
Full text available as:
[img]
Preview
PDF
License: Creative Commons Attribution Non-commercial No Derivatives

Download (334kB) | Preview

Abstract

Non-linear relationships are accommodated in a regression model using smoothing functions. Interaction may occurs between continuous variable, in this case interaction between nonlinear and linear covariate leads to varying coefficent model (VCM), a subclass of generalized additive model. Additive models can be estimated as generalized linear mixed models, after being reparametrized. In this article we show three different type of matrix design for mixed model for VCM, by applying b-spline smoothing functions. An application on real data is provided and model estimates re computed with a Bayesian approach.

Abstract
Document type
Monograph (Working Paper)
Creators
CreatorsAffiliationORCID
Freni Sterrantino, Anna
Keywords
Varying Coefficient models, Generalized linear mixed models, reparametrization, B-spline Modelli a coefficienti variabili, Modelli linearu generaliazzati ad effetti misti, parametrizzazione, B-splinew
Subjects
ISSN
1973-9346
DOI
Deposit date
28 Oct 2013 12:29
Last modified
04 Apr 2016 15:36
URI

Other metadata

Downloads

Downloads

Staff only: View the document

^