[Statlist] Research seminar in statistics December 5th 2014, GSEM, University of Geneva

Eva Cantoni Ev@@C@nton| @end|ng |rom un|ge@ch
Mon Dec 1 13:03:40 CET 2014


Organisers :
E. Cantoni - E. Ronchetti - S. Sperlich - M-P. Victoria-Feser

Friday December 5th, 2014
at 11h15 - Room M 5220, Uni Mail (40, bd du Pont-d'Arve)

Nadège Jacot
Université de Genève

TITLE: Comparison of goodness-of-fit measures in linear mixed-effects models

Abstract:
In linear mixed-effects models, several measures have been proposed in 
order to evaluate the goodness-of-fit of a model, both in the frequentist 
and in the Bayesian framework. We select a large set of these indices 
focusing on those that can be interpreted similarly to the classical 
coeffcient of determination R^2. All indices allow selecting the most 
appropriate model among a set of alternatives, while some additionally 
allow quantifying the quality of the model at hand. First, we briefly 
review and describe the selected goodness-of-fit indices. Then, we 
compute an extensive simulation study, inspired by the varying-intercept 
and varying-slope analysis of radon data in Gelman and Pardoe (2006), to 
evaluate the sensitivity of the selected measures to the values of some 
parameters of the model. Third, we use the indices to select the correct 
model from a series of nested alternative models. Finally we illustrate 
the use of these indices on the popular data of dental measurements of 
Potthoff and Roy (1964).



Visit the website: http://www.stat-center.unige.ch/ressem.html


-- 
Prof. Eva Cantoni
Research Center for Statistics and
      Geneva School of Economics and Management
University of Geneva, Bd du Pont d'Arve 40, CH-1211 Genève 4
http://stat-center.unige.ch/members2/profs/eva-cantoni/




More information about the Statlist mailing list