Bayesian D-optimal design issues for binomial generalized linear model screening designs

Edgar Hassler; Douglas Montgomery; Rachel T. Silvestrini

doi:10.1007/978-3-319-12355-4_20

Bayesian D-optimal design issues for binomial generalized linear model screening designs

Edgar Hassler, Douglas Montgomery, Rachel T. Silvestrini

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

1 Scopus citations

Abstract

Bayesian D-optimal designs have become computationally feasible to construct for simple prior distributions. Some parameter values give rise to models that have little utility to the practitioner for effect screening. For some generalized linear models such as the binomial, inclusion of such models can cause the optimal design to spread out toward the boundary of the design space. This can reduce the D-efficiency of the design over much of the parameter space and result in the Bayesian D-optimal criterion's divergence from the concerns of a practitioner designing a screening experiment.

Original language	English (US)
Title of host publication	Frontiers in Statistical Quality Control 10
Publisher	Kluwer Academic Publishers
Pages	337-353
Number of pages	17
Volume	11
ISBN (Print)	9783319123547
DOIs	https://doi.org/10.1007/978-3-319-12355-4_20
State	Published - 2015
Event	11th International Workshop on Intelligent Statistical Quality Control, 2013 - Sydney, Australia Duration: Aug 20 2013 → Aug 23 2013

Other

Other	11th International Workshop on Intelligent Statistical Quality Control, 2013
Country/Territory	Australia
City	Sydney
Period	8/20/13 → 8/23/13

Keywords

Challenger data set
Confidence intervals
Non-linear designs

ASJC Scopus subject areas

Computer Networks and Communications

Access to Document

10.1007/978-3-319-12355-4_20

Cite this

Hassler, E, Montgomery, D & Silvestrini, RT 2015, Bayesian D-optimal design issues for binomial generalized linear model screening designs. in Frontiers in Statistical Quality Control 10. vol. 11, Kluwer Academic Publishers, pp. 337-353, 11th International Workshop on Intelligent Statistical Quality Control, 2013, Sydney, Australia, 8/20/13. https://doi.org/10.1007/978-3-319-12355-4_20

@inproceedings{92585cbf0d1042199518e0825d5f9f17,

title = "Bayesian D-optimal design issues for binomial generalized linear model screening designs",

abstract = "Bayesian D-optimal designs have become computationally feasible to construct for simple prior distributions. Some parameter values give rise to models that have little utility to the practitioner for effect screening. For some generalized linear models such as the binomial, inclusion of such models can cause the optimal design to spread out toward the boundary of the design space. This can reduce the D-efficiency of the design over much of the parameter space and result in the Bayesian D-optimal criterion's divergence from the concerns of a practitioner designing a screening experiment.",

keywords = "Challenger data set, Confidence intervals, Non-linear designs",

author = "Edgar Hassler and Douglas Montgomery and Silvestrini, {Rachel T.}",

year = "2015",

doi = "10.1007/978-3-319-12355-4_20",

language = "English (US)",

isbn = "9783319123547",

volume = "11",

pages = "337--353",

booktitle = "Frontiers in Statistical Quality Control 10",

publisher = "Kluwer Academic Publishers",

note = "11th International Workshop on Intelligent Statistical Quality Control, 2013 ; Conference date: 20-08-2013 Through 23-08-2013",

}

TY - GEN

T1 - Bayesian D-optimal design issues for binomial generalized linear model screening designs

AU - Hassler, Edgar

AU - Montgomery, Douglas

AU - Silvestrini, Rachel T.

PY - 2015

Y1 - 2015

N2 - Bayesian D-optimal designs have become computationally feasible to construct for simple prior distributions. Some parameter values give rise to models that have little utility to the practitioner for effect screening. For some generalized linear models such as the binomial, inclusion of such models can cause the optimal design to spread out toward the boundary of the design space. This can reduce the D-efficiency of the design over much of the parameter space and result in the Bayesian D-optimal criterion's divergence from the concerns of a practitioner designing a screening experiment.

AB - Bayesian D-optimal designs have become computationally feasible to construct for simple prior distributions. Some parameter values give rise to models that have little utility to the practitioner for effect screening. For some generalized linear models such as the binomial, inclusion of such models can cause the optimal design to spread out toward the boundary of the design space. This can reduce the D-efficiency of the design over much of the parameter space and result in the Bayesian D-optimal criterion's divergence from the concerns of a practitioner designing a screening experiment.

KW - Challenger data set

KW - Confidence intervals

KW - Non-linear designs

UR - http://www.scopus.com/inward/record.url?scp=84946047284&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84946047284&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-12355-4_20

DO - 10.1007/978-3-319-12355-4_20

M3 - Conference contribution

AN - SCOPUS:84946047284

SN - 9783319123547

VL - 11

SP - 337

EP - 353

BT - Frontiers in Statistical Quality Control 10

PB - Kluwer Academic Publishers

T2 - 11th International Workshop on Intelligent Statistical Quality Control, 2013

Y2 - 20 August 2013 through 23 August 2013

ER -

Bayesian D-optimal design issues for binomial generalized linear model screening designs

Abstract

Other

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this