Population parametrization of costly black box models using iterations between SAEM algorithm and kriging

Emmanuel Grenier; Celine Helbert; Violaine Louvet; Adeline Samson; Paul Vigneaux

doi:10.1007/s40314-016-0337-5

Population parametrization of costly black box models using iterations between SAEM algorithm and kriging

Emmanuel Grenier
, Celine Helbert
, Violaine Louvet
, Adeline Samson
, Paul Vigneaux^*

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

2 Citations (Scopus)

Abstract

In this article we focus on parametrization of black box models from repeated measurements among several individuals (population parametrization). We introduce a variant of the SAEM algorithm, called KSAEM algorithm, which couples the standard SAEM algorithm with the dynamic construction of an approximate metamodel. The costly evaluation of the genuine black box is replaced by a kriging step, using a basis of precomputed values, basis which is enlarged during SAEM algorithm to improve the accuracy of the metamodel in regions of interest.

Original language	English
Pages (from-to)	161-173
Number of pages	13
Journal	Computational and Applied Mathematics
Volume	37
Issue number	1
DOIs	https://doi.org/10.1007/s40314-016-0337-5
Publication status	Published - 1 Mar 2018
Externally published	Yes

Keywords

KPP equation
Kriging
Non-linear mixed effect models
Parameters estimation
Partial differential equations
SAEM algorithm

Access to Document

10.1007/s40314-016-0337-5

Cite this

@article{3cb67e080b654297a09b06e0e6e657b8,

title = "Population parametrization of costly black box models using iterations between SAEM algorithm and kriging",

abstract = "In this article we focus on parametrization of black box models from repeated measurements among several individuals (population parametrization). We introduce a variant of the SAEM algorithm, called KSAEM algorithm, which couples the standard SAEM algorithm with the dynamic construction of an approximate metamodel. The costly evaluation of the genuine black box is replaced by a kriging step, using a basis of precomputed values, basis which is enlarged during SAEM algorithm to improve the accuracy of the metamodel in regions of interest.",

keywords = "KPP equation, Kriging, Non-linear mixed effect models, Parameters estimation, Partial differential equations, SAEM algorithm",

author = "Emmanuel Grenier and Celine Helbert and Violaine Louvet and Adeline Samson and Paul Vigneaux",

note = "Publisher Copyright: {\textcopyright} 2016, SBMAC - Sociedade Brasileira de Matem{\'a}tica Aplicada e Computacional.",

year = "2018",

month = mar,

day = "1",

doi = "10.1007/s40314-016-0337-5",

language = "English",

volume = "37",

pages = "161--173",

journal = "Computational and Applied Mathematics",

issn = "2238-3603",

publisher = "Springer Nature",

number = "1",

}

TY - JOUR

T1 - Population parametrization of costly black box models using iterations between SAEM algorithm and kriging

AU - Grenier, Emmanuel

AU - Helbert, Celine

AU - Louvet, Violaine

AU - Samson, Adeline

AU - Vigneaux, Paul

PY - 2018/3/1

Y1 - 2018/3/1

N2 - In this article we focus on parametrization of black box models from repeated measurements among several individuals (population parametrization). We introduce a variant of the SAEM algorithm, called KSAEM algorithm, which couples the standard SAEM algorithm with the dynamic construction of an approximate metamodel. The costly evaluation of the genuine black box is replaced by a kriging step, using a basis of precomputed values, basis which is enlarged during SAEM algorithm to improve the accuracy of the metamodel in regions of interest.

AB - In this article we focus on parametrization of black box models from repeated measurements among several individuals (population parametrization). We introduce a variant of the SAEM algorithm, called KSAEM algorithm, which couples the standard SAEM algorithm with the dynamic construction of an approximate metamodel. The costly evaluation of the genuine black box is replaced by a kriging step, using a basis of precomputed values, basis which is enlarged during SAEM algorithm to improve the accuracy of the metamodel in regions of interest.

KW - KPP equation

KW - Kriging

KW - Non-linear mixed effect models

KW - Parameters estimation

KW - Partial differential equations

KW - SAEM algorithm

UR - https://www.scopus.com/pages/publications/85042549676

U2 - 10.1007/s40314-016-0337-5

DO - 10.1007/s40314-016-0337-5

M3 - Article

AN - SCOPUS:85042549676

SN - 2238-3603

VL - 37

SP - 161

EP - 173

JO - Computational and Applied Mathematics

JF - Computational and Applied Mathematics

IS - 1

ER -

Population parametrization of costly black box models using iterations between SAEM algorithm and kriging

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this