BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Estimation and variable selection in high dimension in nonlinear m
 ixed-effects models
DTSTART:20260127T101500Z
DTEND:20260127T111500Z
DTSTAMP:20260424T050800Z
UID:indico-event-14466@indico.math.cnrs.fr
DESCRIPTION:Speakers: Estelle Kuhn (INRAE - Université Paris Saclay)\n\nJ
 oint work with Antoine Caillebotte (INRAE\, MaIAGE\, GQE-Le Moulon) and Sa
 rah Lemler (CentraleSupelec\, MICS)\n\nIn this work\, we consider nonlinea
 r mixed-effects models including high-dimensional covariates to model indi
 vidual parameters variability. The objective is to identify relevant covar
 iates among a large set under sparsity assumption and to estimate model pa
 rameters. To face the high dimensional setting\, we consider a regularized
  estimator namely the maximum likelihood estimator penalized with the l1-p
 enalty. We rely on the use of the eBIC model choice criterium to select an
  optimal reduced model. Then we estimate the parameters by maximizing the 
 likelihood of the reduced model. We calculate in practice the maximum like
 lihood estimator penalized with the l1-penalty though a weighted proximal 
 stochastic gradient descent algorithm with an adaptive learning rate. This
  choice allows us to consider very general models\, in particular models t
 hat do not belong to the curved exponential family. We demonstrate first i
 n a simple linear toy model through a simulation study the good convergenc
 e properties of this optimization algorithm. We compare then the performan
 ce of the proposed methodology with those of the glmmLasso procedure in a 
 linear mixed-effects model in a simulation study. We illustrate also its p
 erformance in a nonlinear mixed-effects logistic growth model through simu
 lation. We highlight the benefit of the proposed procedure relying on this
  integrated single step approach regarding two others two steps approaches
  for variable selection objective in mixed models. Finally we analyze real
  data of wheat senescence to identify potential relevant markers of this b
 iological process. \n\nhttps://indico.math.cnrs.fr/event/14466/
LOCATION:Salle K. Johnson (1R3\, 1er étage)
URL:https://indico.math.cnrs.fr/event/14466/
END:VEVENT
END:VCALENDAR