SlideShare a Scribd company logo
Copyright © 2008, SAS Institute Inc. All rights reserved.
Practical Design for
Discrete Choice
Experiments
Bradley Jones, SAS Institute
August 13, 2008
Copyright © 2008, SAS Institute Inc. All rights reserved.
DEMA 2008, August 2008,
Cambridge
2
 Respondents indicate the alternative they prefer most in
each choice set
 Alternatives are called profiles
 Each profile is a combination of attribute levels
 Choice sets typically consist of two, three or four profiles
Discrete Choice Experiment Setup
Copyright © 2008, SAS Institute Inc. All rights reserved.
Example: Marketing a new laptop computer
Attributes Levels
Hard Drive 40 GB 80 GB
Speed 1.5 GHz 2.0 GHz
Battery Life 4 hours6 hours
Price $1,000 $1,200 $1,500
DEMA 2008, August 2008,
Cambridge
3
Copyright © 2008, SAS Institute Inc. All rights reserved.
Sample Choice Set
Hard Disk Speed Battery Price
40Gig 1.5GHz 6hours $1,000
40Gig 2.0GHz 4hours $1,500
Check the box for the laptop you prefer.
Profile 1
Profile 2
4DEMA 2008, August 2008,
Cambridge
Copyright © 2008, SAS Institute Inc. All rights reserved.
5
multinomial logit model
based on the random utilities model
where xjs represents the attribute levels and β is the set of
parameter values
probability of choosing alternative j in choice set s
Statistical model
ε′= +js js jsU xβ
=1
option chosen
in choice set
js
ts
js J
t
j e
p
s e
′
′
 
= ÷
  ∑
xβ
xβ
DEMA 2008, August 2008,
Cambridge
Copyright © 2008, SAS Institute Inc. All rights reserved.
D criterion - minimize the determinant of the variance matrix of
the estimators:
( )( )1
det −
M Xβ,
Design optimality criterion
DEMA 2008, August 2008,
Cambridge
6
Equivalently – maximize the determinant of the
information matrix, M.
Copyright © 2008, SAS Institute Inc. All rights reserved.
7
Bayesian optimal designs:
• construct a prior distribution for the parameters
• find design that performs best on average
• Sándor & Wedel (2001, 2002, 2005)
Dependence on the unknown parameter, β
( ) ( ) ( ) ( )( )
1
,
S
s s s s s
s=
′ ′= −∑M Xβ X P β p β p β X
DEMA 2008, August 2008,
Cambridge
Copyright © 2008, SAS Institute Inc. All rights reserved.
Design for Nonlinear Models
To design an informative experiment …..
You need to know something about the response function …..
And about the parameter values.
8DEMA 2008, August 2008,
Cambridge
Copyright © 2008, SAS Institute Inc. All rights reserved.
Bayesian D-Optimal Design
Bayesian ideas are natural to cope with the fact that
the information matrix, M, depends on β.
Chaloner and Larntz (1986) developed a Bayesian D-
Optimality criterion:
Φ(d) = ∫ log det [M(β;d)] p(β) dβ
9DEMA 2008, August 2008,
Cambridge
Copyright © 2008, SAS Institute Inc. All rights reserved.
Computing Bayesian D-Optimal Designs
A major impediment to Bayesian D-optimal design
has been COMPUTATIONAL.
The integral over β can be VERY SLOW.
It must be computed MANY TIMES in the course of
finding an optimal design.
10DEMA 2008, August 2008,
Cambridge
Copyright © 2008, SAS Institute Inc. All rights reserved.
Bayesian Computations
Gotwalt, Jones and Steinberg (2007) use a quadrature
method, due to Mysovskikh.
This method is guaranteed to exactly integrate all
polynomials up to 5th degree and all odd-degree
monomials.
With p parameters, it requires just O(p2) function evaluations.
11DEMA 2008, August 2008,
Cambridge
Copyright © 2008, SAS Institute Inc. All rights reserved.
Mysovskikh quadrature
Assume a normal prior with independence.
• Center the integral about the prior mean.
• Scale each variable by its standard deviation.
• Integrate over distance from the prior mean and, at
each distance, over a spherical shell.
12DEMA 2008, August 2008,
Cambridge
Copyright © 2008, SAS Institute Inc. All rights reserved.
Mysovskikh quadrature continued…
Radial integral: Generalized Gauss-LaGuerre
quadrature, with an extra point at the origin.
Spherical integrals: The Mysovskikh quadrature
scheme.
13DEMA 2008, August 2008,
Cambridge
Copyright © 2008, SAS Institute Inc. All rights reserved.
Spherical integral
A simplex, its edge midpoints on the sphere, and
the inverses of all of these points.
Simplex point weights: p(7-p)/2(p+1)2(p+2).
Mid-point weights: 2(p-1)2/p(p+1)2(p+2).
14DEMA 2008, August 2008,
Cambridge
Copyright © 2008, SAS Institute Inc. All rights reserved.
Simplex point
Mid-point
Each point is both a simplex
point and a mid-point.
All weights equal1/6.
Quadrature points in two dimensions
15DEMA 2008, August 2008,
Cambridge
Copyright © 2008, SAS Institute Inc. All rights reserved.
Study Description
16 Respondents – 8 developers 8 sales & marketing
9 Male 7 Female
2 Surveys with 6 choice sets in each
Respondents were assigned randomly to surveys
blocked by job function
16DEMA 2008, August 2008,
Cambridge
Copyright © 2008, SAS Institute Inc. All rights reserved.
Software Demonstration
17DEMA 2008, August 2008,
Cambridge
Copyright © 2008, SAS Institute Inc. All rights reserved.
Conclusions
1) Discrete Choice Conjoint Experiments require design
methods for nonlinear models.
2) D-Optimal Bayesian designs reduce the dependence
of the design on the unknown parameters.
3) New quadrature methods make computation of these
designs much faster.
4) Commercial software makes carrying out such
studies simple and efficient.
18DEMA 2008, August 2008,
Cambridge
Copyright © 2008, SAS Institute Inc. All rights reserved.
References
Atkinson, A. C. and Donev, A. N. (1992). Optimum Experimental Designs, Oxford U.K.: Clarendon Press.
Cassity C.R., (1965) “Abscissas, Coefficients, and Error Term for the Generalized Gauss-Laguerre Quadrature Formula Using the Zero
Ordinate,” Mathematics ofComputation, 19, 287-296.
Chaloner, K. and Verdinelli, I. (1995). Bayesian experimental design: a re-view, Statistical Science 10: 273-304.
Grossmann, H., Holling, H. and Schwabe, R. (2002). Advances in optimum experimental design for conjoint analysis and discrete
choice models, in Advances in Econometrics, Econometric Models in Marketing, Vol. 16, Franses, P. H. and Montgomery, A. L.,
eds. Amsterdam: JAI Press, 93-117.
Gotwalt, C., Jones, B. and Steinberg, D. (2009) Fast Computation of Designs Robust to Parameter Uncertainty for Nonlinear Settings
accepted at Technometrics.
Huber, J. and Zwerina, K. (1996). The importance of utility balance in efficient choice designs, Journal of Marketing Research 33: 307-
317.
McFadden, D. (1974). Conditional logit analysis of qualitative choice behavior, in Frontiers in Econometrics, Zarembka, P., ed. New
York: Academic Press, 105-142.
Meyer, R. K. and Nachtsheim, C. J. (1995). The coordinate-exchange algorithm for constructing exact optimal experimental designs,
Technometrics 37: 60-69.
Monahan, J. and Genz, A. (1997). Spherical-radial integration rules for Bayesian computation, Journal of the American Statistical
Association 92: 664-674.
Sandor, Z. and Wedel, M. (2001). Designing conjoint choice experiments using managers' prior beliefs, Journal of Marketing Research
38: 430-444.
19DEMA 2008, August 2008,
Cambridge

More Related Content

Viewers also liked

20091127 Ec Gaal Estali 2
20091127 Ec Gaal Estali 220091127 Ec Gaal Estali 2
20091127 Ec Gaal Estali 2
gaalnorb
 
А что будет с командой? (Yandex & SPB Software)
А что будет с командой? (Yandex & SPB Software)А что будет с командой? (Yandex & SPB Software)
А что будет с командой? (Yandex & SPB Software)
Ксения Артеменко
 
Almuerzo bachilleres
Almuerzo bachilleresAlmuerzo bachilleres
Almuerzo bachilleres
ASOPADRESINJUV
 
Berkarir di dunia telekomunikasi
Berkarir di dunia telekomunikasiBerkarir di dunia telekomunikasi
Berkarir di dunia telekomunikasi
datux
 
The saga of satyam
The saga of satyamThe saga of satyam
The saga of satyam
Janki Kalaria
 
Raissa powerpoin[1] 1
Raissa powerpoin[1] 1Raissa powerpoin[1] 1
Raissa powerpoin[1] 1
ecsmedia
 
my horror film presentation
my horror film presentationmy horror film presentation
my horror film presentation
ecsmedia
 
Media (horror movie storyboard by Aamir
Media (horror movie storyboard  by AamirMedia (horror movie storyboard  by Aamir
Media (horror movie storyboard by Aamir
ecsmedia
 
Pinar's Media Presentation.
Pinar's Media Presentation.Pinar's Media Presentation.
Pinar's Media Presentation.
ecsmedia
 
Alippt[1]
Alippt[1]Alippt[1]
Alippt[1]
ecsmedia
 
ESD/CRT Discussion Notes
ESD/CRT Discussion NotesESD/CRT Discussion Notes
ESD/CRT Discussion Notes
bluebuilding
 
Guardians of the Leads Webinar
Guardians of the Leads WebinarGuardians of the Leads Webinar
Guardians of the Leads Webinar
MarketStar Corp
 
linkCharge
linkChargelinkCharge
linkChargepaySto
 
Tilve maria Fête du albariño cambados- Tilve Maria
Tilve maria Fête du albariño cambados- Tilve MariaTilve maria Fête du albariño cambados- Tilve Maria
Tilve maria Fête du albariño cambados- Tilve Maria
kedougou
 
Dead man walking
Dead man walkingDead man walking
Dead man walking
ecsmedia
 
Jordel's shot type presentation
Jordel's shot type presentationJordel's shot type presentation
Jordel's shot type presentation
ecsmedia
 
Jopr29
Jopr29Jopr29

Viewers also liked (20)

20091127 Ec Gaal Estali 2
20091127 Ec Gaal Estali 220091127 Ec Gaal Estali 2
20091127 Ec Gaal Estali 2
 
А что будет с командой? (Yandex & SPB Software)
А что будет с командой? (Yandex & SPB Software)А что будет с командой? (Yandex & SPB Software)
А что будет с командой? (Yandex & SPB Software)
 
Almuerzo bachilleres
Almuerzo bachilleresAlmuerzo bachilleres
Almuerzo bachilleres
 
Berkarir di dunia telekomunikasi
Berkarir di dunia telekomunikasiBerkarir di dunia telekomunikasi
Berkarir di dunia telekomunikasi
 
The saga of satyam
The saga of satyamThe saga of satyam
The saga of satyam
 
Raissa powerpoin[1] 1
Raissa powerpoin[1] 1Raissa powerpoin[1] 1
Raissa powerpoin[1] 1
 
my horror film presentation
my horror film presentationmy horror film presentation
my horror film presentation
 
Media (horror movie storyboard by Aamir
Media (horror movie storyboard  by AamirMedia (horror movie storyboard  by Aamir
Media (horror movie storyboard by Aamir
 
Pinar's Media Presentation.
Pinar's Media Presentation.Pinar's Media Presentation.
Pinar's Media Presentation.
 
Alippt[1]
Alippt[1]Alippt[1]
Alippt[1]
 
ESD/CRT Discussion Notes
ESD/CRT Discussion NotesESD/CRT Discussion Notes
ESD/CRT Discussion Notes
 
Guardians of the Leads Webinar
Guardians of the Leads WebinarGuardians of the Leads Webinar
Guardians of the Leads Webinar
 
linkCharge
linkChargelinkCharge
linkCharge
 
Tilve maria Fête du albariño cambados- Tilve Maria
Tilve maria Fête du albariño cambados- Tilve MariaTilve maria Fête du albariño cambados- Tilve Maria
Tilve maria Fête du albariño cambados- Tilve Maria
 
Dead man walking
Dead man walkingDead man walking
Dead man walking
 
Presentation3
Presentation3Presentation3
Presentation3
 
Autonome voertuigen
Autonome voertuigenAutonome voertuigen
Autonome voertuigen
 
Jordel's shot type presentation
Jordel's shot type presentationJordel's shot type presentation
Jordel's shot type presentation
 
Jopr29
Jopr29Jopr29
Jopr29
 
Gta artwork
Gta artworkGta artwork
Gta artwork
 

Similar to upload test1

Software tools, crystal descriptors, and machine learning applied to material...
Software tools, crystal descriptors, and machine learning applied to material...Software tools, crystal descriptors, and machine learning applied to material...
Software tools, crystal descriptors, and machine learning applied to material...
Anubhav Jain
 
Accelerated life testing
Accelerated life testingAccelerated life testing
Accelerated life testing
Steven Li
 
Srikanta Mishra
Srikanta MishraSrikanta Mishra
Vldb14
Vldb14Vldb14
Vldb14
hdbtracker
 
Six Sigma Mechanical Tolerance Analysis 1
Six Sigma Mechanical Tolerance Analysis 1Six Sigma Mechanical Tolerance Analysis 1
Six Sigma Mechanical Tolerance Analysis 1
David Panek
 
Staad pro
Staad proStaad pro
Big Data + Big Sim: Query Processing over Unstructured CFD Models
Big Data + Big Sim: Query Processing over Unstructured CFD ModelsBig Data + Big Sim: Query Processing over Unstructured CFD Models
Big Data + Big Sim: Query Processing over Unstructured CFD Models
University of Washington
 
Jorge Silva, Sr. Research Statistician Developer, SAS at MLconf ATL - 9/18/15
Jorge Silva, Sr. Research Statistician Developer, SAS at MLconf ATL - 9/18/15Jorge Silva, Sr. Research Statistician Developer, SAS at MLconf ATL - 9/18/15
Jorge Silva, Sr. Research Statistician Developer, SAS at MLconf ATL - 9/18/15
MLconf
 
Chap 8. Optimization for training deep models
Chap 8. Optimization for training deep modelsChap 8. Optimization for training deep models
Chap 8. Optimization for training deep models
Young-Geun Choi
 
Hyperparameter optimization with approximate gradient
Hyperparameter optimization with approximate gradientHyperparameter optimization with approximate gradient
Hyperparameter optimization with approximate gradient
Fabian Pedregosa
 
Automated Machine Learning Applied to Diverse Materials Design Problems
Automated Machine Learning Applied to Diverse Materials Design ProblemsAutomated Machine Learning Applied to Diverse Materials Design Problems
Automated Machine Learning Applied to Diverse Materials Design Problems
Anubhav Jain
 
theory of computation lecture 01
theory of computation lecture 01theory of computation lecture 01
theory of computation lecture 01
8threspecter
 
Sparse Isotropic Hashing
Sparse Isotropic HashingSparse Isotropic Hashing
Sparse Isotropic Hashing
Ikuro Sato
 
Real Time Geodemographics
Real Time GeodemographicsReal Time Geodemographics
Real Time Geodemographics
Dr Muhammad Adnan
 
Licentiate Defense Slide
Licentiate Defense SlideLicentiate Defense Slide
Licentiate Defense Slide
Rerngvit Yanggratoke
 
Efficient evaluation of flatness error from Coordinate Measurement Data using...
Efficient evaluation of flatness error from Coordinate Measurement Data using...Efficient evaluation of flatness error from Coordinate Measurement Data using...
Efficient evaluation of flatness error from Coordinate Measurement Data using...
Ali Shahed
 
Multi-criteria Decision Analysis for Customization of Estimation by Analogy M...
Multi-criteria Decision Analysis for Customization of Estimation by Analogy M...Multi-criteria Decision Analysis for Customization of Estimation by Analogy M...
Multi-criteria Decision Analysis for Customization of Estimation by Analogy M...
gregoryg
 
A simulation based multi-objective design optimization of electronic packages...
A simulation based multi-objective design optimization of electronic packages...A simulation based multi-objective design optimization of electronic packages...
A simulation based multi-objective design optimization of electronic packages...
Phuong Dx
 
IRJET- Review of Existing Methods in K-Means Clustering Algorithm
IRJET- Review of Existing Methods in K-Means Clustering AlgorithmIRJET- Review of Existing Methods in K-Means Clustering Algorithm
IRJET- Review of Existing Methods in K-Means Clustering Algorithm
IRJET Journal
 
A Framework and Infrastructure for Uncertainty Quantification and Management ...
A Framework and Infrastructure for Uncertainty Quantification and Management ...A Framework and Infrastructure for Uncertainty Quantification and Management ...
A Framework and Infrastructure for Uncertainty Quantification and Management ...
aimsnist
 

Similar to upload test1 (20)

Software tools, crystal descriptors, and machine learning applied to material...
Software tools, crystal descriptors, and machine learning applied to material...Software tools, crystal descriptors, and machine learning applied to material...
Software tools, crystal descriptors, and machine learning applied to material...
 
Accelerated life testing
Accelerated life testingAccelerated life testing
Accelerated life testing
 
Srikanta Mishra
Srikanta MishraSrikanta Mishra
Srikanta Mishra
 
Vldb14
Vldb14Vldb14
Vldb14
 
Six Sigma Mechanical Tolerance Analysis 1
Six Sigma Mechanical Tolerance Analysis 1Six Sigma Mechanical Tolerance Analysis 1
Six Sigma Mechanical Tolerance Analysis 1
 
Staad pro
Staad proStaad pro
Staad pro
 
Big Data + Big Sim: Query Processing over Unstructured CFD Models
Big Data + Big Sim: Query Processing over Unstructured CFD ModelsBig Data + Big Sim: Query Processing over Unstructured CFD Models
Big Data + Big Sim: Query Processing over Unstructured CFD Models
 
Jorge Silva, Sr. Research Statistician Developer, SAS at MLconf ATL - 9/18/15
Jorge Silva, Sr. Research Statistician Developer, SAS at MLconf ATL - 9/18/15Jorge Silva, Sr. Research Statistician Developer, SAS at MLconf ATL - 9/18/15
Jorge Silva, Sr. Research Statistician Developer, SAS at MLconf ATL - 9/18/15
 
Chap 8. Optimization for training deep models
Chap 8. Optimization for training deep modelsChap 8. Optimization for training deep models
Chap 8. Optimization for training deep models
 
Hyperparameter optimization with approximate gradient
Hyperparameter optimization with approximate gradientHyperparameter optimization with approximate gradient
Hyperparameter optimization with approximate gradient
 
Automated Machine Learning Applied to Diverse Materials Design Problems
Automated Machine Learning Applied to Diverse Materials Design ProblemsAutomated Machine Learning Applied to Diverse Materials Design Problems
Automated Machine Learning Applied to Diverse Materials Design Problems
 
theory of computation lecture 01
theory of computation lecture 01theory of computation lecture 01
theory of computation lecture 01
 
Sparse Isotropic Hashing
Sparse Isotropic HashingSparse Isotropic Hashing
Sparse Isotropic Hashing
 
Real Time Geodemographics
Real Time GeodemographicsReal Time Geodemographics
Real Time Geodemographics
 
Licentiate Defense Slide
Licentiate Defense SlideLicentiate Defense Slide
Licentiate Defense Slide
 
Efficient evaluation of flatness error from Coordinate Measurement Data using...
Efficient evaluation of flatness error from Coordinate Measurement Data using...Efficient evaluation of flatness error from Coordinate Measurement Data using...
Efficient evaluation of flatness error from Coordinate Measurement Data using...
 
Multi-criteria Decision Analysis for Customization of Estimation by Analogy M...
Multi-criteria Decision Analysis for Customization of Estimation by Analogy M...Multi-criteria Decision Analysis for Customization of Estimation by Analogy M...
Multi-criteria Decision Analysis for Customization of Estimation by Analogy M...
 
A simulation based multi-objective design optimization of electronic packages...
A simulation based multi-objective design optimization of electronic packages...A simulation based multi-objective design optimization of electronic packages...
A simulation based multi-objective design optimization of electronic packages...
 
IRJET- Review of Existing Methods in K-Means Clustering Algorithm
IRJET- Review of Existing Methods in K-Means Clustering AlgorithmIRJET- Review of Existing Methods in K-Means Clustering Algorithm
IRJET- Review of Existing Methods in K-Means Clustering Algorithm
 
A Framework and Infrastructure for Uncertainty Quantification and Management ...
A Framework and Infrastructure for Uncertainty Quantification and Management ...A Framework and Infrastructure for Uncertainty Quantification and Management ...
A Framework and Infrastructure for Uncertainty Quantification and Management ...
 

More from sahilsahoo85

sigproc-sp.pdf
sigproc-sp.pdfsigproc-sp.pdf
sigproc-sp.pdf
sahilsahoo85
 
CHANGELOG.txt
CHANGELOG.txtCHANGELOG.txt
CHANGELOG.txt
sahilsahoo85
 
testing_7.ppt
testing_7.ppttesting_7.ppt
testing_7.ppt
sahilsahoo85
 
testing_6.ppt
testing_6.ppttesting_6.ppt
testing_6.ppt
sahilsahoo85
 
testing_5.ppt
testing_5.ppttesting_5.ppt
testing_5.ppt
sahilsahoo85
 
testing_5.ppt
testing_5.ppttesting_5.ppt
testing_5.ppt
sahilsahoo85
 
testing_5.ppt
testing_5.ppttesting_5.ppt
testing_5.ppt
sahilsahoo85
 
testing_5.ppt
testing_5.ppttesting_5.ppt
testing_5.ppt
sahilsahoo85
 
testing_5.ppt
testing_5.ppttesting_5.ppt
testing_5.ppt
sahilsahoo85
 
testing_4.ppt
testing_4.ppttesting_4.ppt
testing_4.ppt
sahilsahoo85
 
testing_3.ppt
testing_3.ppttesting_3.ppt
testing_3.ppt
sahilsahoo85
 
testing_2.ppt
testing_2.ppttesting_2.ppt
testing_2.ppt
sahilsahoo85
 
questionnaries-theme.tpl_.php_.txt
questionnaries-theme.tpl_.php_.txtquestionnaries-theme.tpl_.php_.txt
questionnaries-theme.tpl_.php_.txt
sahilsahoo85
 
ccccccc.ppt
ccccccc.pptccccccc.ppt
ccccccc.ppt
sahilsahoo85
 

More from sahilsahoo85 (20)

sigproc-sp.pdf
sigproc-sp.pdfsigproc-sp.pdf
sigproc-sp.pdf
 
CHANGELOG.txt
CHANGELOG.txtCHANGELOG.txt
CHANGELOG.txt
 
testing_7.ppt
testing_7.ppttesting_7.ppt
testing_7.ppt
 
testing_6.ppt
testing_6.ppttesting_6.ppt
testing_6.ppt
 
testing_5.ppt
testing_5.ppttesting_5.ppt
testing_5.ppt
 
testing_5.ppt
testing_5.ppttesting_5.ppt
testing_5.ppt
 
testing_5.ppt
testing_5.ppttesting_5.ppt
testing_5.ppt
 
testing_5.ppt
testing_5.ppttesting_5.ppt
testing_5.ppt
 
testing_5.ppt
testing_5.ppttesting_5.ppt
testing_5.ppt
 
testing_4.ppt
testing_4.ppttesting_4.ppt
testing_4.ppt
 
testing_3.ppt
testing_3.ppttesting_3.ppt
testing_3.ppt
 
testing_2.ppt
testing_2.ppttesting_2.ppt
testing_2.ppt
 
testing_1.ppt
testing_1.ppttesting_1.ppt
testing_1.ppt
 
testing_0.ppt
testing_0.ppttesting_0.ppt
testing_0.ppt
 
testing_0.ppt
testing_0.ppttesting_0.ppt
testing_0.ppt
 
questionnaries-theme.tpl_.php_.txt
questionnaries-theme.tpl_.php_.txtquestionnaries-theme.tpl_.php_.txt
questionnaries-theme.tpl_.php_.txt
 
ccccccc.ppt
ccccccc.pptccccccc.ppt
ccccccc.ppt
 
testing.ppt
testing.ppttesting.ppt
testing.ppt
 
testing_0.ppt
testing_0.ppttesting_0.ppt
testing_0.ppt
 
upload test2
upload test2upload test2
upload test2
 

Recently uploaded

Dandelion Hashtable: beyond billion requests per second on a commodity server
Dandelion Hashtable: beyond billion requests per second on a commodity serverDandelion Hashtable: beyond billion requests per second on a commodity server
Dandelion Hashtable: beyond billion requests per second on a commodity server
Antonios Katsarakis
 
Columbus Data & Analytics Wednesdays - June 2024
Columbus Data & Analytics Wednesdays - June 2024Columbus Data & Analytics Wednesdays - June 2024
Columbus Data & Analytics Wednesdays - June 2024
Jason Packer
 
Leveraging the Graph for Clinical Trials and Standards
Leveraging the Graph for Clinical Trials and StandardsLeveraging the Graph for Clinical Trials and Standards
Leveraging the Graph for Clinical Trials and Standards
Neo4j
 
High performance Serverless Java on AWS- GoTo Amsterdam 2024
High performance Serverless Java on AWS- GoTo Amsterdam 2024High performance Serverless Java on AWS- GoTo Amsterdam 2024
High performance Serverless Java on AWS- GoTo Amsterdam 2024
Vadym Kazulkin
 
9 CEO's who hit $100m ARR Share Their Top Growth Tactics Nathan Latka, Founde...
9 CEO's who hit $100m ARR Share Their Top Growth Tactics Nathan Latka, Founde...9 CEO's who hit $100m ARR Share Their Top Growth Tactics Nathan Latka, Founde...
9 CEO's who hit $100m ARR Share Their Top Growth Tactics Nathan Latka, Founde...
saastr
 
Apps Break Data
Apps Break DataApps Break Data
Apps Break Data
Ivo Velitchkov
 
"$10 thousand per minute of downtime: architecture, queues, streaming and fin...
"$10 thousand per minute of downtime: architecture, queues, streaming and fin..."$10 thousand per minute of downtime: architecture, queues, streaming and fin...
"$10 thousand per minute of downtime: architecture, queues, streaming and fin...
Fwdays
 
5th LF Energy Power Grid Model Meet-up Slides
5th LF Energy Power Grid Model Meet-up Slides5th LF Energy Power Grid Model Meet-up Slides
5th LF Energy Power Grid Model Meet-up Slides
DanBrown980551
 
zkStudyClub - LatticeFold: A Lattice-based Folding Scheme and its Application...
zkStudyClub - LatticeFold: A Lattice-based Folding Scheme and its Application...zkStudyClub - LatticeFold: A Lattice-based Folding Scheme and its Application...
zkStudyClub - LatticeFold: A Lattice-based Folding Scheme and its Application...
Alex Pruden
 
The Microsoft 365 Migration Tutorial For Beginner.pptx
The Microsoft 365 Migration Tutorial For Beginner.pptxThe Microsoft 365 Migration Tutorial For Beginner.pptx
The Microsoft 365 Migration Tutorial For Beginner.pptx
operationspcvita
 
JavaLand 2024: Application Development Green Masterplan
JavaLand 2024: Application Development Green MasterplanJavaLand 2024: Application Development Green Masterplan
JavaLand 2024: Application Development Green Masterplan
Miro Wengner
 
Session 1 - Intro to Robotic Process Automation.pdf
Session 1 - Intro to Robotic Process Automation.pdfSession 1 - Intro to Robotic Process Automation.pdf
Session 1 - Intro to Robotic Process Automation.pdf
UiPathCommunity
 
Must Know Postgres Extension for DBA and Developer during Migration
Must Know Postgres Extension for DBA and Developer during MigrationMust Know Postgres Extension for DBA and Developer during Migration
Must Know Postgres Extension for DBA and Developer during Migration
Mydbops
 
Taking AI to the Next Level in Manufacturing.pdf
Taking AI to the Next Level in Manufacturing.pdfTaking AI to the Next Level in Manufacturing.pdf
Taking AI to the Next Level in Manufacturing.pdf
ssuserfac0301
 
GNSS spoofing via SDR (Criptored Talks 2024)
GNSS spoofing via SDR (Criptored Talks 2024)GNSS spoofing via SDR (Criptored Talks 2024)
GNSS spoofing via SDR (Criptored Talks 2024)
Javier Junquera
 
Main news related to the CCS TSI 2023 (2023/1695)
Main news related to the CCS TSI 2023 (2023/1695)Main news related to the CCS TSI 2023 (2023/1695)
Main news related to the CCS TSI 2023 (2023/1695)
Jakub Marek
 
Northern Engraving | Modern Metal Trim, Nameplates and Appliance Panels
Northern Engraving | Modern Metal Trim, Nameplates and Appliance PanelsNorthern Engraving | Modern Metal Trim, Nameplates and Appliance Panels
Northern Engraving | Modern Metal Trim, Nameplates and Appliance Panels
Northern Engraving
 
Your One-Stop Shop for Python Success: Top 10 US Python Development Providers
Your One-Stop Shop for Python Success: Top 10 US Python Development ProvidersYour One-Stop Shop for Python Success: Top 10 US Python Development Providers
Your One-Stop Shop for Python Success: Top 10 US Python Development Providers
akankshawande
 
GraphRAG for LifeSciences Hands-On with the Clinical Knowledge Graph
GraphRAG for LifeSciences Hands-On with the Clinical Knowledge GraphGraphRAG for LifeSciences Hands-On with the Clinical Knowledge Graph
GraphRAG for LifeSciences Hands-On with the Clinical Knowledge Graph
Neo4j
 
Fueling AI with Great Data with Airbyte Webinar
Fueling AI with Great Data with Airbyte WebinarFueling AI with Great Data with Airbyte Webinar
Fueling AI with Great Data with Airbyte Webinar
Zilliz
 

Recently uploaded (20)

Dandelion Hashtable: beyond billion requests per second on a commodity server
Dandelion Hashtable: beyond billion requests per second on a commodity serverDandelion Hashtable: beyond billion requests per second on a commodity server
Dandelion Hashtable: beyond billion requests per second on a commodity server
 
Columbus Data & Analytics Wednesdays - June 2024
Columbus Data & Analytics Wednesdays - June 2024Columbus Data & Analytics Wednesdays - June 2024
Columbus Data & Analytics Wednesdays - June 2024
 
Leveraging the Graph for Clinical Trials and Standards
Leveraging the Graph for Clinical Trials and StandardsLeveraging the Graph for Clinical Trials and Standards
Leveraging the Graph for Clinical Trials and Standards
 
High performance Serverless Java on AWS- GoTo Amsterdam 2024
High performance Serverless Java on AWS- GoTo Amsterdam 2024High performance Serverless Java on AWS- GoTo Amsterdam 2024
High performance Serverless Java on AWS- GoTo Amsterdam 2024
 
9 CEO's who hit $100m ARR Share Their Top Growth Tactics Nathan Latka, Founde...
9 CEO's who hit $100m ARR Share Their Top Growth Tactics Nathan Latka, Founde...9 CEO's who hit $100m ARR Share Their Top Growth Tactics Nathan Latka, Founde...
9 CEO's who hit $100m ARR Share Their Top Growth Tactics Nathan Latka, Founde...
 
Apps Break Data
Apps Break DataApps Break Data
Apps Break Data
 
"$10 thousand per minute of downtime: architecture, queues, streaming and fin...
"$10 thousand per minute of downtime: architecture, queues, streaming and fin..."$10 thousand per minute of downtime: architecture, queues, streaming and fin...
"$10 thousand per minute of downtime: architecture, queues, streaming and fin...
 
5th LF Energy Power Grid Model Meet-up Slides
5th LF Energy Power Grid Model Meet-up Slides5th LF Energy Power Grid Model Meet-up Slides
5th LF Energy Power Grid Model Meet-up Slides
 
zkStudyClub - LatticeFold: A Lattice-based Folding Scheme and its Application...
zkStudyClub - LatticeFold: A Lattice-based Folding Scheme and its Application...zkStudyClub - LatticeFold: A Lattice-based Folding Scheme and its Application...
zkStudyClub - LatticeFold: A Lattice-based Folding Scheme and its Application...
 
The Microsoft 365 Migration Tutorial For Beginner.pptx
The Microsoft 365 Migration Tutorial For Beginner.pptxThe Microsoft 365 Migration Tutorial For Beginner.pptx
The Microsoft 365 Migration Tutorial For Beginner.pptx
 
JavaLand 2024: Application Development Green Masterplan
JavaLand 2024: Application Development Green MasterplanJavaLand 2024: Application Development Green Masterplan
JavaLand 2024: Application Development Green Masterplan
 
Session 1 - Intro to Robotic Process Automation.pdf
Session 1 - Intro to Robotic Process Automation.pdfSession 1 - Intro to Robotic Process Automation.pdf
Session 1 - Intro to Robotic Process Automation.pdf
 
Must Know Postgres Extension for DBA and Developer during Migration
Must Know Postgres Extension for DBA and Developer during MigrationMust Know Postgres Extension for DBA and Developer during Migration
Must Know Postgres Extension for DBA and Developer during Migration
 
Taking AI to the Next Level in Manufacturing.pdf
Taking AI to the Next Level in Manufacturing.pdfTaking AI to the Next Level in Manufacturing.pdf
Taking AI to the Next Level in Manufacturing.pdf
 
GNSS spoofing via SDR (Criptored Talks 2024)
GNSS spoofing via SDR (Criptored Talks 2024)GNSS spoofing via SDR (Criptored Talks 2024)
GNSS spoofing via SDR (Criptored Talks 2024)
 
Main news related to the CCS TSI 2023 (2023/1695)
Main news related to the CCS TSI 2023 (2023/1695)Main news related to the CCS TSI 2023 (2023/1695)
Main news related to the CCS TSI 2023 (2023/1695)
 
Northern Engraving | Modern Metal Trim, Nameplates and Appliance Panels
Northern Engraving | Modern Metal Trim, Nameplates and Appliance PanelsNorthern Engraving | Modern Metal Trim, Nameplates and Appliance Panels
Northern Engraving | Modern Metal Trim, Nameplates and Appliance Panels
 
Your One-Stop Shop for Python Success: Top 10 US Python Development Providers
Your One-Stop Shop for Python Success: Top 10 US Python Development ProvidersYour One-Stop Shop for Python Success: Top 10 US Python Development Providers
Your One-Stop Shop for Python Success: Top 10 US Python Development Providers
 
GraphRAG for LifeSciences Hands-On with the Clinical Knowledge Graph
GraphRAG for LifeSciences Hands-On with the Clinical Knowledge GraphGraphRAG for LifeSciences Hands-On with the Clinical Knowledge Graph
GraphRAG for LifeSciences Hands-On with the Clinical Knowledge Graph
 
Fueling AI with Great Data with Airbyte Webinar
Fueling AI with Great Data with Airbyte WebinarFueling AI with Great Data with Airbyte Webinar
Fueling AI with Great Data with Airbyte Webinar
 

upload test1

  • 1. Copyright © 2008, SAS Institute Inc. All rights reserved. Practical Design for Discrete Choice Experiments Bradley Jones, SAS Institute August 13, 2008
  • 2. Copyright © 2008, SAS Institute Inc. All rights reserved. DEMA 2008, August 2008, Cambridge 2  Respondents indicate the alternative they prefer most in each choice set  Alternatives are called profiles  Each profile is a combination of attribute levels  Choice sets typically consist of two, three or four profiles Discrete Choice Experiment Setup
  • 3. Copyright © 2008, SAS Institute Inc. All rights reserved. Example: Marketing a new laptop computer Attributes Levels Hard Drive 40 GB 80 GB Speed 1.5 GHz 2.0 GHz Battery Life 4 hours6 hours Price $1,000 $1,200 $1,500 DEMA 2008, August 2008, Cambridge 3
  • 4. Copyright © 2008, SAS Institute Inc. All rights reserved. Sample Choice Set Hard Disk Speed Battery Price 40Gig 1.5GHz 6hours $1,000 40Gig 2.0GHz 4hours $1,500 Check the box for the laptop you prefer. Profile 1 Profile 2 4DEMA 2008, August 2008, Cambridge
  • 5. Copyright © 2008, SAS Institute Inc. All rights reserved. 5 multinomial logit model based on the random utilities model where xjs represents the attribute levels and β is the set of parameter values probability of choosing alternative j in choice set s Statistical model ε′= +js js jsU xβ =1 option chosen in choice set js ts js J t j e p s e ′ ′   = ÷   ∑ xβ xβ DEMA 2008, August 2008, Cambridge
  • 6. Copyright © 2008, SAS Institute Inc. All rights reserved. D criterion - minimize the determinant of the variance matrix of the estimators: ( )( )1 det − M Xβ, Design optimality criterion DEMA 2008, August 2008, Cambridge 6 Equivalently – maximize the determinant of the information matrix, M.
  • 7. Copyright © 2008, SAS Institute Inc. All rights reserved. 7 Bayesian optimal designs: • construct a prior distribution for the parameters • find design that performs best on average • Sándor & Wedel (2001, 2002, 2005) Dependence on the unknown parameter, β ( ) ( ) ( ) ( )( ) 1 , S s s s s s s= ′ ′= −∑M Xβ X P β p β p β X DEMA 2008, August 2008, Cambridge
  • 8. Copyright © 2008, SAS Institute Inc. All rights reserved. Design for Nonlinear Models To design an informative experiment ….. You need to know something about the response function ….. And about the parameter values. 8DEMA 2008, August 2008, Cambridge
  • 9. Copyright © 2008, SAS Institute Inc. All rights reserved. Bayesian D-Optimal Design Bayesian ideas are natural to cope with the fact that the information matrix, M, depends on β. Chaloner and Larntz (1986) developed a Bayesian D- Optimality criterion: Φ(d) = ∫ log det [M(β;d)] p(β) dβ 9DEMA 2008, August 2008, Cambridge
  • 10. Copyright © 2008, SAS Institute Inc. All rights reserved. Computing Bayesian D-Optimal Designs A major impediment to Bayesian D-optimal design has been COMPUTATIONAL. The integral over β can be VERY SLOW. It must be computed MANY TIMES in the course of finding an optimal design. 10DEMA 2008, August 2008, Cambridge
  • 11. Copyright © 2008, SAS Institute Inc. All rights reserved. Bayesian Computations Gotwalt, Jones and Steinberg (2007) use a quadrature method, due to Mysovskikh. This method is guaranteed to exactly integrate all polynomials up to 5th degree and all odd-degree monomials. With p parameters, it requires just O(p2) function evaluations. 11DEMA 2008, August 2008, Cambridge
  • 12. Copyright © 2008, SAS Institute Inc. All rights reserved. Mysovskikh quadrature Assume a normal prior with independence. • Center the integral about the prior mean. • Scale each variable by its standard deviation. • Integrate over distance from the prior mean and, at each distance, over a spherical shell. 12DEMA 2008, August 2008, Cambridge
  • 13. Copyright © 2008, SAS Institute Inc. All rights reserved. Mysovskikh quadrature continued… Radial integral: Generalized Gauss-LaGuerre quadrature, with an extra point at the origin. Spherical integrals: The Mysovskikh quadrature scheme. 13DEMA 2008, August 2008, Cambridge
  • 14. Copyright © 2008, SAS Institute Inc. All rights reserved. Spherical integral A simplex, its edge midpoints on the sphere, and the inverses of all of these points. Simplex point weights: p(7-p)/2(p+1)2(p+2). Mid-point weights: 2(p-1)2/p(p+1)2(p+2). 14DEMA 2008, August 2008, Cambridge
  • 15. Copyright © 2008, SAS Institute Inc. All rights reserved. Simplex point Mid-point Each point is both a simplex point and a mid-point. All weights equal1/6. Quadrature points in two dimensions 15DEMA 2008, August 2008, Cambridge
  • 16. Copyright © 2008, SAS Institute Inc. All rights reserved. Study Description 16 Respondents – 8 developers 8 sales & marketing 9 Male 7 Female 2 Surveys with 6 choice sets in each Respondents were assigned randomly to surveys blocked by job function 16DEMA 2008, August 2008, Cambridge
  • 17. Copyright © 2008, SAS Institute Inc. All rights reserved. Software Demonstration 17DEMA 2008, August 2008, Cambridge
  • 18. Copyright © 2008, SAS Institute Inc. All rights reserved. Conclusions 1) Discrete Choice Conjoint Experiments require design methods for nonlinear models. 2) D-Optimal Bayesian designs reduce the dependence of the design on the unknown parameters. 3) New quadrature methods make computation of these designs much faster. 4) Commercial software makes carrying out such studies simple and efficient. 18DEMA 2008, August 2008, Cambridge
  • 19. Copyright © 2008, SAS Institute Inc. All rights reserved. References Atkinson, A. C. and Donev, A. N. (1992). Optimum Experimental Designs, Oxford U.K.: Clarendon Press. Cassity C.R., (1965) “Abscissas, Coefficients, and Error Term for the Generalized Gauss-Laguerre Quadrature Formula Using the Zero Ordinate,” Mathematics ofComputation, 19, 287-296. Chaloner, K. and Verdinelli, I. (1995). Bayesian experimental design: a re-view, Statistical Science 10: 273-304. Grossmann, H., Holling, H. and Schwabe, R. (2002). Advances in optimum experimental design for conjoint analysis and discrete choice models, in Advances in Econometrics, Econometric Models in Marketing, Vol. 16, Franses, P. H. and Montgomery, A. L., eds. Amsterdam: JAI Press, 93-117. Gotwalt, C., Jones, B. and Steinberg, D. (2009) Fast Computation of Designs Robust to Parameter Uncertainty for Nonlinear Settings accepted at Technometrics. Huber, J. and Zwerina, K. (1996). The importance of utility balance in efficient choice designs, Journal of Marketing Research 33: 307- 317. McFadden, D. (1974). Conditional logit analysis of qualitative choice behavior, in Frontiers in Econometrics, Zarembka, P., ed. New York: Academic Press, 105-142. Meyer, R. K. and Nachtsheim, C. J. (1995). The coordinate-exchange algorithm for constructing exact optimal experimental designs, Technometrics 37: 60-69. Monahan, J. and Genz, A. (1997). Spherical-radial integration rules for Bayesian computation, Journal of the American Statistical Association 92: 664-674. Sandor, Z. and Wedel, M. (2001). Designing conjoint choice experiments using managers' prior beliefs, Journal of Marketing Research 38: 430-444. 19DEMA 2008, August 2008, Cambridge