SlideShare une entreprise Scribd logo
1  sur  41
Télécharger pour lire hors ligne
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
.
......
Challenges with the Use of Cross-validation for
Comparing Structured Models
Wei Wang
joint work with Andrew Gelman
Department of Statistics, Columbia University
April 13, 2013
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Overview
...1 Multilevel Models
...2 Decision-eoretic Model Assessment Framework
...3 Data and Model
...4 Results
Overview
...1 Multilevel Models
...2 Decision-eoretic Model Assessment Framework
...3 Data and Model
...4 Results
. . . . . .
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Bayesian Interpretation of Multilevel Models
Multilevel Models have long been proposed to handle data with
group structures, e.g., longitudinal study with multiple obs. for
each participant, national survey with various demographic and
geographic variables.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Bayesian Interpretation of Multilevel Models
Multilevel Models have long been proposed to handle data with
group structures, e.g., longitudinal study with multiple obs. for
each participant, national survey with various demographic and
geographic variables.
From a Bayesian point of view, what Multilevel Modeling does is
to partially pool the estimates through a prior, as opposed to
doing separate analysis for each group (no pooling) or analyzing
the data as if there is no group structure (complete pooling).
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Multilevel Models for Deeply Nested Data Structure
Our substantive interest is survey data with deeply nested
structures resulting from various categorical
demographic-geographic variables, e.g., state, income, education,
ethnicity et al.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Multilevel Models for Deeply Nested Data Structure
Our substantive interest is survey data with deeply nested
structures resulting from various categorical
demographic-geographic variables, e.g., state, income, education,
ethnicity et al.
One typical conundrum is how many interactions between those
demographic-geographic variables to include in the model.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
ree Prototypes of Models
In the simple case of two predictors, the three prototypes of models are
shown below. e response yi is binary.
Complete Pooling model
Eyij ∼ g−1
(µij)
µij = µ0 + ai + bj
No Pooling model
Eyij ∼ g−1
(µij)
µij = µ0 + ai + bj + rij
Partial Pooling model
Eyij ∼ g−1
(µij)
µij = µ0 + ai + bj + γij
γ ∼ Φ(·)
Overview
...1 Multilevel Models
...2 Decision-eoretic Model Assessment Framework
...3 Data and Model
...4 Results
. . . . . .
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
True model, Pseudo-true model and Actual Belief model
We assume there is a true underlying model pt(·), from which the
observations (both available and future observations) come from.
While acknowledging the fact that the true distribution is never
accessible, some researchers propose basing the discussion on a
rich enough Actual Belief Model), which supposedly fully re ects
the uncertainty of future data. (Bernardo and Smith 1994)
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
M-closed, M-completed and M-open views
In M-closed view, it is assumed that the true model is included
in a enumerable collection of models, and the Actual Belief
Model is the Bayesian Model Averaging predictive distribution.
In M-completed view, the Actual Belief Model p(˜y|D, M) is
considered to be the best available description of the uncertainty
of future data.
In M-open view, the correct speci cation of the Actual Belief
Model is avoided and the strategy is to generate Monte Carlo
samples from it, such as sample re-use methods.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
A Decision-eoretical Framework
We de ne a loss function l(˜y, aM), which is the loss incurred
from our inferential action aM, based on a model M, in face of
future observation ˜y.
en the predictive loss from our inferential action aM is
Lp(pt
, M, D, l) = Ept(˜y)l(˜y, aM) =
∫
l(˜y, aM)pt
(˜y)d˜y
It is oen convenient and theoretically desirable to use the whole
posterior predictive distribution as aM and the log loss as l(·, ·).
Lpred(pt,M,D)=Ept [− log p(˜y|D,M)]=−
∫
pt(˜y) log p(˜y|D,M)d˜y
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Decision-eoretic Framework Cont'd
For Model Selection task, from a pool of candidate models
{Mk : k ∈ K}, we should select the model that minimizes the
expected predictive loss.
min
Mk:k∈K
−
∫
pt
(˜y) log p(˜y|D, M)d˜y
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Decision-eoretic Framework Cont'd
For Model Selection task, from a pool of candidate models
{Mk : k ∈ K}, we should select the model that minimizes the
expected predictive loss.
min
Mk:k∈K
−
∫
pt
(˜y) log p(˜y|D, M)d˜y
For Model Assessment task of a particular model M, we look at
the Kullback-Leibler divergence between the true model and the
posterior predictive distribution. We call it the predictive error.
Err(pt
, M, D) = −
∫
pt
(˜y) log p(˜y|D, M)d˜y +
∫
pt
(˜y) log pt
(˜y)d˜y
= KL(p(·|D, M); pt
(·))
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Estimating Expected Predictive Loss
e central obstacle of getting the Expected Predicitve Loss is
that we don't know the true distribution pt(·).
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Estimating Expected Predictive Loss
e central obstacle of getting the Expected Predicitve Loss is
that we don't know the true distribution pt(·).
A M-closed or M-completed view will substitute the true
distribution with a reference distribution.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Estimating Expected Predictive Loss
e central obstacle of getting the Expected Predicitve Loss is
that we don't know the true distribution pt(·).
A M-closed or M-completed view will substitute the true
distribution with a reference distribution.
From a M-open view, plug in available sample gives us the
Training Loss, which has a downward bias, since we used the
sample twice.
Ltraining(M, D) = −
1
n
n∑
i=1
log p(yi|D, M)
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Estimating Expected Predictive Loss
e central obstacle of getting the Expected Predicitve Loss is
that we don't know the true distribution pt(·).
A M-closed or M-completed view will substitute the true
distribution with a reference distribution.
From a M-open view, plug in available sample gives us the
Training Loss, which has a downward bias, since we used the
sample twice.
Ltraining(M, D) = −
1
n
n∑
i=1
log p(yi|D, M)
ere exist two approaches to get an unbiased estimate of
Predictive Loss: Bias Correction which leads to various
Information Criteria; Held-out Practices which lead to
Leave-one-out Cross Validation and k-fold Cross Validation.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Estimation Methods
ere is a long list of variants of Information Criteria,
AIC/BIC/DIC/TIC/NIC/WAIC et al.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Estimation Methods
ere is a long list of variants of Information Criteria,
AIC/BIC/DIC/TIC/NIC/WAIC et al.
LOO Cross Validation has been shown to be asymptotically
equivalent to AIC/WAIC. But the computational burden is huge.
e Importance Sampling method introduces new problem of
the reliability of the importance weights.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Estimation Methods
ere is a long list of variants of Information Criteria,
AIC/BIC/DIC/TIC/NIC/WAIC et al.
LOO Cross Validation has been shown to be asymptotically
equivalent to AIC/WAIC. But the computational burden is huge.
e Importance Sampling method introduces new problem of
the reliability of the importance weights.
We are using the computationally convenient k-fold cross
validation, in which the data set is randomly partitioned into k
parts, and in each fold, one part is used as the testing set while
the rest serve as the training set.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
k-fold Cross Validation
en the k-fold Cross Validation estimate of the Predictive Loss
is given by
LCV(M, D) = −
K∑
k=1
∑
i∈testk
log p(yi|Dk
, M) = −
N∑
i=1
log p(yi|D(i)
, M)
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
k-fold Cross Validation
en the k-fold Cross Validation estimate of the Predictive Loss
is given by
LCV(M, D) = −
K∑
k=1
∑
i∈testk
log p(yi|Dk
, M) = −
N∑
i=1
log p(yi|D(i)
, M)
To estimate the Predictive Error, we still need an estimate of the
Entropy of the true distribution. We can use the training loss of
the saturated model as a surrogate.
−
∫
pt(˜y) log pt(˜y)d˜y = −
1
n
n∑
i=1
log p(˙yi|D, Msaturated)
Overview
...1 Multilevel Models
...2 Decision-eoretic Model Assessment Framework
...3 Data and Model
...4 Results
. . . . . .
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Data Set
Cooperative Congressional Election Survey 2006
N=30,000
71 social and political response outcomes
Deeply nested demographic variables, e.g., state, inc, edu, ethn,
gender et al.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Data Set Cont'd
Figure: A sample of the questions in CCES 2006 survey.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Model Setup
For demonstration, we only consider two demographic variables,
state and income, together with their interaction. e responses
are all yes-no binary outcomes.
Complete Pooling
πj1j2
= logit−1
(
βstt
j1
+ βinc
j2
)
No Pooling
πj1j2
= logit−1
(
βstt
j1
+ βinc
j2
+ βstt*inc
j1j2
)
Partial Pooling
πj1j2
= logit−1
(
βstt
j1
+ βinc
j2
+ βstt*inc
j1j2
)
βstt*inc
j1j2
∼ Φ(·)
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
k-fold Cross Validation Estimate
Due to computational constraints, we are using Maximum A
Posteriori plug-in estimate instead of full Bayesian estimate.
p(˜y|D, M) ≈ p(˜y|ˆπij(D), M)
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
k-fold Cross Validation Estimate
Due to computational constraints, we are using Maximum A
Posteriori plug-in estimate instead of full Bayesian estimate.
p(˜y|D, M) ≈ p(˜y|ˆπij(D), M)
en under the aforementioned setup, the Cross Validation
estimate of the Predictive Loss is
LCV(M,D)=− 1
N
∑K
k=1
∑
l∈testk
log p(yl|Dk,M)
=− 1
N
∑K
k=1
∑
i,j[y
testk
ij log ˆπij(Dtraink )+(n
testk
ij −y
testk
ij ) log(1−ˆπij(Dtraink ))]
=− 1
N
∑
i,j
∑K
k=1[log ˆπij(Dtraink )y
testk
ij +log(1−ˆπij(Dtraink ))(n
testk
ij −y
testk
ij )]
=− 1
N
∑
i,j
[
log ˆπij(Dtrain)yij+log(1−ˆπij(Dtrain))(nij−yij)
]
=−
∑
i,j
nij
N
[
log ˆπij(Dtrain)˜πij+log(1−ˆπij(Dtrain))(1−˜πij)
]
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Calibration of Improvement
Let's suppose we only have one cell, with true proportion .4, and
the good model gives a posterior estimate of log proportion at
roughly log(0.41), and the lesser model gives a estimate of
log(0.44) or log(0.38).
en the Predictive Loss under the good model is
−[.4 ∗ log(.41) + .6 ∗ log(.59)] = 0.67322, and under the two
lesser models is −[.4 ∗ log(.44) + .6 ∗ log(.56)] = 0.67386 and
−[.4 ∗ log(.38) + .6 ∗ log(.62)] = 0.67628. We can see the
improvement of the Predictive Loss is between 0.0006 to 0.003.
Also, the lower bound is given by
−[.4 ∗ log(.4) + .6 ∗ log(.6)] = 0.67301, so the Predictive Error
of the good model is about 0.0002.
Overview
...1 Multilevel Models
...2 Decision-eoretic Model Assessment Framework
...3 Data and Model
...4 Results
. . . . . .
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Cross Validation Results on All Outcomes
Responses (ordered by the lower bound)
EstimatedPredictiveError
0.01
0.02
0.03
0.04
0.05
10 20 30 40 50 60 70
models
complete pooling
partial pooling
no pooling
Figure: Measure of t (Estimated Predictive Error) for all response outcomes
in CCES 2006 survey data. Responses are ordered by the lower bound
(training loss of the saturated model). No Pooling model gives very bad t,
while Predictive Error of Partial Pooling is dominated by Complete Pooling,
but the differences seem small.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Compare Partial Pooling and Complete Pooling
In the previous gure, apparently No Pooling is doing very badly,
but the differences between Partial Pooling and Complete
Pooling seem small. We need to further calibrate them.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Compare Partial Pooling and Complete Pooling
In the previous gure, apparently No Pooling is doing very badly,
but the differences between Partial Pooling and Complete
Pooling seem small. We need to further calibrate them.
e summary of the differences between Partial Pooling and
Complete Pooling for all the outcomes is
Min. 1st Qu. Median Mean 3rd Qu. Max.
-0.0003405 0.0001821 0.0003827 0.0006041 0.0005630 0.0053770
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Compare Partial Pooling and Complete Pooling
In the previous gure, apparently No Pooling is doing very badly,
but the differences between Partial Pooling and Complete
Pooling seem small. We need to further calibrate them.
e summary of the differences between Partial Pooling and
Complete Pooling for all the outcomes is
Min. 1st Qu. Median Mean 3rd Qu. Max.
-0.0003405 0.0001821 0.0003827 0.0006041 0.0005630 0.0053770
We can see that the improvement in terms of the Predictive Loss
indeed corresponds to some meaningful improvement in
prediction accuracy.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Simulations Based on Real Data
We want to explore how the structure of the multilevel models
affects the dynamics of the performance of different models.
Speci cally, we are interested in total sample size and how
balanced the cells are in terms of cell size.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Simulations Based on Real Data
We want to explore how the structure of the multilevel models
affects the dynamics of the performance of different models.
Speci cally, we are interested in total sample size and how
balanced the cells are in terms of cell size.
We generated simulated data sets based on the real data set, i.e.,
we use the estimated from the Multilevel model t of the real data
sets and enlarge the total sample size by 2, 3 and 4 times, either
keeping the original relative proportions (highly unequal) of
different cells or making the proportions roughly equal.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Simulation Results: Total Sample Size
Responses (ordered by the lower bound)
EstimatedPredictiveError
0.002
0.003
0.004
0.005
10 20 30 40 50 60 70
models
complete pooling
partial pooling
no pooling
Responses (ordered by the lower bound)
EstimatedPredictiveError
0.0020
0.0025
0.0030
0.0035
0.0040
0.0045
10 20 30 40 50 60 70
models
complete pooling
partial pooling
no pooling
Responses (ordered by the lower bound)
EstimatedPredictiveError
0.002
0.003
0.004
0.005
0.006
10 20 30 40 50 60 70
models
complete pooling
partial pooling
no pooling
Figure: Estimated Predictive Error of all response outcomes for
``augmented'' data sets.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Simulation Results: Total Sample Size on House Rep Vote
sample size
EstimatedPredictiveError
0.002
0.004
0.006
0.008
0.010
0.012
0.014
50000 100000 150000 200000
models
complete pooling
partial pooling
no pooling
Figure: Predictive Error of the three models as sample size grows. e
outcome under consideration is the Republican vote in the House election.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Simulation Results: Balancedness of the Structure
Responses (ordered by the lower bound)
EstimatedPredictiveError
0.010
0.015
0.020
0.025
0.030
10 20 30 40 50 60 70
models
complete pooling
partial pooling
no pooling
Figure: Measure of t (Predictive Error) for all responses, ordered by lower
bound. e data set is simulated from real data set, and has the same sample
size in total as the real data set, but keeping all demographic-geographic cells
balanced.
. . . . . .
Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results
Conclusions
Cross-validation is not a very sensitive instrument in comparing
multilevel models.
Careful calibrations are needed for better understanding of the
results.
We also explored how different aspects of the data set structure
affect the margin of improvement.

Contenu connexe

Tendances

Eugen Zaharescu-PROJECT STATEMENT-Morphological Medical Image Indexing and Cl...
Eugen Zaharescu-PROJECT STATEMENT-Morphological Medical Image Indexing and Cl...Eugen Zaharescu-PROJECT STATEMENT-Morphological Medical Image Indexing and Cl...
Eugen Zaharescu-PROJECT STATEMENT-Morphological Medical Image Indexing and Cl...
Eugen Zaharescu
 

Tendances (18)

Conv xg
Conv xgConv xg
Conv xg
 
Msb12e ppt ch11
Msb12e ppt ch11Msb12e ppt ch11
Msb12e ppt ch11
 
Comparison on PCA ICA and LDA in Face Recognition
Comparison on PCA ICA and LDA in Face RecognitionComparison on PCA ICA and LDA in Face Recognition
Comparison on PCA ICA and LDA in Face Recognition
 
PREDICTIVE EVALUATION OF THE STOCK PORTFOLIO PERFORMANCE USING FUZZY CMEANS A...
PREDICTIVE EVALUATION OF THE STOCK PORTFOLIO PERFORMANCE USING FUZZY CMEANS A...PREDICTIVE EVALUATION OF THE STOCK PORTFOLIO PERFORMANCE USING FUZZY CMEANS A...
PREDICTIVE EVALUATION OF THE STOCK PORTFOLIO PERFORMANCE USING FUZZY CMEANS A...
 
IRJET-Handwritten Digit Classification using Machine Learning Models
IRJET-Handwritten Digit Classification using Machine Learning ModelsIRJET-Handwritten Digit Classification using Machine Learning Models
IRJET-Handwritten Digit Classification using Machine Learning Models
 
Machine learning for computer vision part 2
Machine learning for computer vision part 2Machine learning for computer vision part 2
Machine learning for computer vision part 2
 
Assess curr mathg03
Assess curr mathg03Assess curr mathg03
Assess curr mathg03
 
Machine learning in science and industry — day 1
Machine learning in science and industry — day 1Machine learning in science and industry — day 1
Machine learning in science and industry — day 1
 
Assess curr mathg07
Assess curr mathg07Assess curr mathg07
Assess curr mathg07
 
Conditional mixture model for modeling attributed dyadic data
Conditional mixture model for modeling attributed dyadic dataConditional mixture model for modeling attributed dyadic data
Conditional mixture model for modeling attributed dyadic data
 
Binary Class and Multi Class Strategies for Machine Learning
Binary Class and Multi Class Strategies for Machine LearningBinary Class and Multi Class Strategies for Machine Learning
Binary Class and Multi Class Strategies for Machine Learning
 
Eugen Zaharescu-PROJECT STATEMENT-Morphological Medical Image Indexing and Cl...
Eugen Zaharescu-PROJECT STATEMENT-Morphological Medical Image Indexing and Cl...Eugen Zaharescu-PROJECT STATEMENT-Morphological Medical Image Indexing and Cl...
Eugen Zaharescu-PROJECT STATEMENT-Morphological Medical Image Indexing and Cl...
 
High-Dimensional Methods: Examples for Inference on Structural Effects
High-Dimensional Methods: Examples for Inference on Structural EffectsHigh-Dimensional Methods: Examples for Inference on Structural Effects
High-Dimensional Methods: Examples for Inference on Structural Effects
 
Machine learning in science and industry — day 2
Machine learning in science and industry — day 2Machine learning in science and industry — day 2
Machine learning in science and industry — day 2
 
Assess curr mathg08
Assess curr mathg08Assess curr mathg08
Assess curr mathg08
 
Ga
GaGa
Ga
 
Fundementals of Machine Learning and Deep Learning
Fundementals of Machine Learning and Deep Learning Fundementals of Machine Learning and Deep Learning
Fundementals of Machine Learning and Deep Learning
 
A New Approach for Ranking Shadowed Fuzzy Numbers and its Application
A New Approach for Ranking Shadowed Fuzzy Numbers and its ApplicationA New Approach for Ranking Shadowed Fuzzy Numbers and its Application
A New Approach for Ranking Shadowed Fuzzy Numbers and its Application
 

Similaire à Minghui Conference Cross-Validation Talk

Scaling Multinomial Logistic Regression via Hybrid Parallelism
Scaling Multinomial Logistic Regression via Hybrid ParallelismScaling Multinomial Logistic Regression via Hybrid Parallelism
Scaling Multinomial Logistic Regression via Hybrid Parallelism
Parameswaran Raman
 
Probability density estimation using Product of Conditional Experts
Probability density estimation using Product of Conditional ExpertsProbability density estimation using Product of Conditional Experts
Probability density estimation using Product of Conditional Experts
Chirag Gupta
 
Intro to Model Selection
Intro to Model SelectionIntro to Model Selection
Intro to Model Selection
chenhm
 
Declarative data analysis
Declarative data analysisDeclarative data analysis
Declarative data analysis
South West Data Meetup
 
Machine Learning Tutorial Part - 2 | Machine Learning Tutorial For Beginners ...
Machine Learning Tutorial Part - 2 | Machine Learning Tutorial For Beginners ...Machine Learning Tutorial Part - 2 | Machine Learning Tutorial For Beginners ...
Machine Learning Tutorial Part - 2 | Machine Learning Tutorial For Beginners ...
Simplilearn
 
Course Project for Coursera Practical Machine Learning
Course Project for Coursera Practical Machine LearningCourse Project for Coursera Practical Machine Learning
Course Project for Coursera Practical Machine Learning
John Edward Slough II
 

Similaire à Minghui Conference Cross-Validation Talk (20)

Scaling Multinomial Logistic Regression via Hybrid Parallelism
Scaling Multinomial Logistic Regression via Hybrid ParallelismScaling Multinomial Logistic Regression via Hybrid Parallelism
Scaling Multinomial Logistic Regression via Hybrid Parallelism
 
G. Barcaroli, The use of machine learning in official statistics
G. Barcaroli, The use of machine learning in official statisticsG. Barcaroli, The use of machine learning in official statistics
G. Barcaroli, The use of machine learning in official statistics
 
[update] Introductory Parts of the Book "Dive into Deep Learning"
[update] Introductory Parts of the Book "Dive into Deep Learning"[update] Introductory Parts of the Book "Dive into Deep Learning"
[update] Introductory Parts of the Book "Dive into Deep Learning"
 
Probability density estimation using Product of Conditional Experts
Probability density estimation using Product of Conditional ExpertsProbability density estimation using Product of Conditional Experts
Probability density estimation using Product of Conditional Experts
 
Intro to Model Selection
Intro to Model SelectionIntro to Model Selection
Intro to Model Selection
 
Machine learning ppt unit one syllabuspptx
Machine learning ppt unit one syllabuspptxMachine learning ppt unit one syllabuspptx
Machine learning ppt unit one syllabuspptx
 
THE IMPLICATION OF STATISTICAL ANALYSIS AND FEATURE ENGINEERING FOR MODEL BUI...
THE IMPLICATION OF STATISTICAL ANALYSIS AND FEATURE ENGINEERING FOR MODEL BUI...THE IMPLICATION OF STATISTICAL ANALYSIS AND FEATURE ENGINEERING FOR MODEL BUI...
THE IMPLICATION OF STATISTICAL ANALYSIS AND FEATURE ENGINEERING FOR MODEL BUI...
 
THE IMPLICATION OF STATISTICAL ANALYSIS AND FEATURE ENGINEERING FOR MODEL BUI...
THE IMPLICATION OF STATISTICAL ANALYSIS AND FEATURE ENGINEERING FOR MODEL BUI...THE IMPLICATION OF STATISTICAL ANALYSIS AND FEATURE ENGINEERING FOR MODEL BUI...
THE IMPLICATION OF STATISTICAL ANALYSIS AND FEATURE ENGINEERING FOR MODEL BUI...
 
Multimodal Biometrics Recognition by Dimensionality Diminution Method
Multimodal Biometrics Recognition by Dimensionality Diminution MethodMultimodal Biometrics Recognition by Dimensionality Diminution Method
Multimodal Biometrics Recognition by Dimensionality Diminution Method
 
Tech meetup Data Driven - Codemotion
Tech meetup Data Driven - Codemotion Tech meetup Data Driven - Codemotion
Tech meetup Data Driven - Codemotion
 
Building and deploying analytics
Building and deploying analyticsBuilding and deploying analytics
Building and deploying analytics
 
Declarative data analysis
Declarative data analysisDeclarative data analysis
Declarative data analysis
 
2.8 accuracy and ensemble methods
2.8 accuracy and ensemble methods2.8 accuracy and ensemble methods
2.8 accuracy and ensemble methods
 
Machine Learning Tutorial Part - 2 | Machine Learning Tutorial For Beginners ...
Machine Learning Tutorial Part - 2 | Machine Learning Tutorial For Beginners ...Machine Learning Tutorial Part - 2 | Machine Learning Tutorial For Beginners ...
Machine Learning Tutorial Part - 2 | Machine Learning Tutorial For Beginners ...
 
Model Uncertainty
Model UncertaintyModel Uncertainty
Model Uncertainty
 
AlgorithmsModelsNov13.pptx
AlgorithmsModelsNov13.pptxAlgorithmsModelsNov13.pptx
AlgorithmsModelsNov13.pptx
 
MyStataLab Assignment Help
MyStataLab Assignment HelpMyStataLab Assignment Help
MyStataLab Assignment Help
 
IMAGE CLASSIFICATION USING DIFFERENT CLASSICAL APPROACHES
IMAGE CLASSIFICATION USING DIFFERENT CLASSICAL APPROACHESIMAGE CLASSIFICATION USING DIFFERENT CLASSICAL APPROACHES
IMAGE CLASSIFICATION USING DIFFERENT CLASSICAL APPROACHES
 
Module 2: Machine Learning Deep Dive
Module 2:  Machine Learning Deep DiveModule 2:  Machine Learning Deep Dive
Module 2: Machine Learning Deep Dive
 
Course Project for Coursera Practical Machine Learning
Course Project for Coursera Practical Machine LearningCourse Project for Coursera Practical Machine Learning
Course Project for Coursera Practical Machine Learning
 

Dernier

Call Girls From Pari Chowk Greater Noida ❤️8448577510 ⊹Best Escorts Service I...
Call Girls From Pari Chowk Greater Noida ❤️8448577510 ⊹Best Escorts Service I...Call Girls From Pari Chowk Greater Noida ❤️8448577510 ⊹Best Escorts Service I...
Call Girls From Pari Chowk Greater Noida ❤️8448577510 ⊹Best Escorts Service I...
lizamodels9
 
Chandigarh Escorts Service 📞8868886958📞 Just📲 Call Nihal Chandigarh Call Girl...
Chandigarh Escorts Service 📞8868886958📞 Just📲 Call Nihal Chandigarh Call Girl...Chandigarh Escorts Service 📞8868886958📞 Just📲 Call Nihal Chandigarh Call Girl...
Chandigarh Escorts Service 📞8868886958📞 Just📲 Call Nihal Chandigarh Call Girl...
Sheetaleventcompany
 
FULL ENJOY Call Girls In Majnu Ka Tilla, Delhi Contact Us 8377877756
FULL ENJOY Call Girls In Majnu Ka Tilla, Delhi Contact Us 8377877756FULL ENJOY Call Girls In Majnu Ka Tilla, Delhi Contact Us 8377877756
FULL ENJOY Call Girls In Majnu Ka Tilla, Delhi Contact Us 8377877756
dollysharma2066
 
Call Girls in Delhi, Escort Service Available 24x7 in Delhi 959961-/-3876
Call Girls in Delhi, Escort Service Available 24x7 in Delhi 959961-/-3876Call Girls in Delhi, Escort Service Available 24x7 in Delhi 959961-/-3876
Call Girls in Delhi, Escort Service Available 24x7 in Delhi 959961-/-3876
dlhescort
 
FULL ENJOY Call Girls In Mahipalpur Delhi Contact Us 8377877756
FULL ENJOY Call Girls In Mahipalpur Delhi Contact Us 8377877756FULL ENJOY Call Girls In Mahipalpur Delhi Contact Us 8377877756
FULL ENJOY Call Girls In Mahipalpur Delhi Contact Us 8377877756
dollysharma2066
 

Dernier (20)

Malegaon Call Girls Service ☎ ️82500–77686 ☎️ Enjoy 24/7 Escort Service
Malegaon Call Girls Service ☎ ️82500–77686 ☎️ Enjoy 24/7 Escort ServiceMalegaon Call Girls Service ☎ ️82500–77686 ☎️ Enjoy 24/7 Escort Service
Malegaon Call Girls Service ☎ ️82500–77686 ☎️ Enjoy 24/7 Escort Service
 
Call Girls From Pari Chowk Greater Noida ❤️8448577510 ⊹Best Escorts Service I...
Call Girls From Pari Chowk Greater Noida ❤️8448577510 ⊹Best Escorts Service I...Call Girls From Pari Chowk Greater Noida ❤️8448577510 ⊹Best Escorts Service I...
Call Girls From Pari Chowk Greater Noida ❤️8448577510 ⊹Best Escorts Service I...
 
It will be International Nurses' Day on 12 May
It will be International Nurses' Day on 12 MayIt will be International Nurses' Day on 12 May
It will be International Nurses' Day on 12 May
 
Chandigarh Escorts Service 📞8868886958📞 Just📲 Call Nihal Chandigarh Call Girl...
Chandigarh Escorts Service 📞8868886958📞 Just📲 Call Nihal Chandigarh Call Girl...Chandigarh Escorts Service 📞8868886958📞 Just📲 Call Nihal Chandigarh Call Girl...
Chandigarh Escorts Service 📞8868886958📞 Just📲 Call Nihal Chandigarh Call Girl...
 
RSA Conference Exhibitor List 2024 - Exhibitors Data
RSA Conference Exhibitor List 2024 - Exhibitors DataRSA Conference Exhibitor List 2024 - Exhibitors Data
RSA Conference Exhibitor List 2024 - Exhibitors Data
 
Eluru Call Girls Service ☎ ️93326-06886 ❤️‍🔥 Enjoy 24/7 Escort Service
Eluru Call Girls Service ☎ ️93326-06886 ❤️‍🔥 Enjoy 24/7 Escort ServiceEluru Call Girls Service ☎ ️93326-06886 ❤️‍🔥 Enjoy 24/7 Escort Service
Eluru Call Girls Service ☎ ️93326-06886 ❤️‍🔥 Enjoy 24/7 Escort Service
 
Call Girls Service In Old Town Dubai ((0551707352)) Old Town Dubai Call Girl ...
Call Girls Service In Old Town Dubai ((0551707352)) Old Town Dubai Call Girl ...Call Girls Service In Old Town Dubai ((0551707352)) Old Town Dubai Call Girl ...
Call Girls Service In Old Town Dubai ((0551707352)) Old Town Dubai Call Girl ...
 
Falcon's Invoice Discounting: Your Path to Prosperity
Falcon's Invoice Discounting: Your Path to ProsperityFalcon's Invoice Discounting: Your Path to Prosperity
Falcon's Invoice Discounting: Your Path to Prosperity
 
FULL ENJOY Call Girls In Majnu Ka Tilla, Delhi Contact Us 8377877756
FULL ENJOY Call Girls In Majnu Ka Tilla, Delhi Contact Us 8377877756FULL ENJOY Call Girls In Majnu Ka Tilla, Delhi Contact Us 8377877756
FULL ENJOY Call Girls In Majnu Ka Tilla, Delhi Contact Us 8377877756
 
Call Girls Zirakpur👧 Book Now📱7837612180 📞👉Call Girl Service In Zirakpur No A...
Call Girls Zirakpur👧 Book Now📱7837612180 📞👉Call Girl Service In Zirakpur No A...Call Girls Zirakpur👧 Book Now📱7837612180 📞👉Call Girl Service In Zirakpur No A...
Call Girls Zirakpur👧 Book Now📱7837612180 📞👉Call Girl Service In Zirakpur No A...
 
Falcon Invoice Discounting: The best investment platform in india for investors
Falcon Invoice Discounting: The best investment platform in india for investorsFalcon Invoice Discounting: The best investment platform in india for investors
Falcon Invoice Discounting: The best investment platform in india for investors
 
👉Chandigarh Call Girls 👉9878799926👉Just Call👉Chandigarh Call Girl In Chandiga...
👉Chandigarh Call Girls 👉9878799926👉Just Call👉Chandigarh Call Girl In Chandiga...👉Chandigarh Call Girls 👉9878799926👉Just Call👉Chandigarh Call Girl In Chandiga...
👉Chandigarh Call Girls 👉9878799926👉Just Call👉Chandigarh Call Girl In Chandiga...
 
Mysore Call Girls 8617370543 WhatsApp Number 24x7 Best Services
Mysore Call Girls 8617370543 WhatsApp Number 24x7 Best ServicesMysore Call Girls 8617370543 WhatsApp Number 24x7 Best Services
Mysore Call Girls 8617370543 WhatsApp Number 24x7 Best Services
 
Call Girls in Delhi, Escort Service Available 24x7 in Delhi 959961-/-3876
Call Girls in Delhi, Escort Service Available 24x7 in Delhi 959961-/-3876Call Girls in Delhi, Escort Service Available 24x7 in Delhi 959961-/-3876
Call Girls in Delhi, Escort Service Available 24x7 in Delhi 959961-/-3876
 
FULL ENJOY Call Girls In Mahipalpur Delhi Contact Us 8377877756
FULL ENJOY Call Girls In Mahipalpur Delhi Contact Us 8377877756FULL ENJOY Call Girls In Mahipalpur Delhi Contact Us 8377877756
FULL ENJOY Call Girls In Mahipalpur Delhi Contact Us 8377877756
 
JAYNAGAR CALL GIRL IN 98274*61493 ❤CALL GIRLS IN ESCORT SERVICE❤CALL GIRL
JAYNAGAR CALL GIRL IN 98274*61493 ❤CALL GIRLS IN ESCORT SERVICE❤CALL GIRLJAYNAGAR CALL GIRL IN 98274*61493 ❤CALL GIRLS IN ESCORT SERVICE❤CALL GIRL
JAYNAGAR CALL GIRL IN 98274*61493 ❤CALL GIRLS IN ESCORT SERVICE❤CALL GIRL
 
Value Proposition canvas- Customer needs and pains
Value Proposition canvas- Customer needs and painsValue Proposition canvas- Customer needs and pains
Value Proposition canvas- Customer needs and pains
 
Business Model Canvas (BMC)- A new venture concept
Business Model Canvas (BMC)-  A new venture conceptBusiness Model Canvas (BMC)-  A new venture concept
Business Model Canvas (BMC)- A new venture concept
 
How to Get Started in Social Media for Art League City
How to Get Started in Social Media for Art League CityHow to Get Started in Social Media for Art League City
How to Get Started in Social Media for Art League City
 
Uneak White's Personal Brand Exploration Presentation
Uneak White's Personal Brand Exploration PresentationUneak White's Personal Brand Exploration Presentation
Uneak White's Personal Brand Exploration Presentation
 

Minghui Conference Cross-Validation Talk

  • 1. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results . ...... Challenges with the Use of Cross-validation for Comparing Structured Models Wei Wang joint work with Andrew Gelman Department of Statistics, Columbia University April 13, 2013
  • 2. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Overview ...1 Multilevel Models ...2 Decision-eoretic Model Assessment Framework ...3 Data and Model ...4 Results
  • 3. Overview ...1 Multilevel Models ...2 Decision-eoretic Model Assessment Framework ...3 Data and Model ...4 Results . . . . . .
  • 4. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Bayesian Interpretation of Multilevel Models Multilevel Models have long been proposed to handle data with group structures, e.g., longitudinal study with multiple obs. for each participant, national survey with various demographic and geographic variables.
  • 5. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Bayesian Interpretation of Multilevel Models Multilevel Models have long been proposed to handle data with group structures, e.g., longitudinal study with multiple obs. for each participant, national survey with various demographic and geographic variables. From a Bayesian point of view, what Multilevel Modeling does is to partially pool the estimates through a prior, as opposed to doing separate analysis for each group (no pooling) or analyzing the data as if there is no group structure (complete pooling).
  • 6. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Multilevel Models for Deeply Nested Data Structure Our substantive interest is survey data with deeply nested structures resulting from various categorical demographic-geographic variables, e.g., state, income, education, ethnicity et al.
  • 7. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Multilevel Models for Deeply Nested Data Structure Our substantive interest is survey data with deeply nested structures resulting from various categorical demographic-geographic variables, e.g., state, income, education, ethnicity et al. One typical conundrum is how many interactions between those demographic-geographic variables to include in the model.
  • 8. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results ree Prototypes of Models In the simple case of two predictors, the three prototypes of models are shown below. e response yi is binary. Complete Pooling model Eyij ∼ g−1 (µij) µij = µ0 + ai + bj No Pooling model Eyij ∼ g−1 (µij) µij = µ0 + ai + bj + rij Partial Pooling model Eyij ∼ g−1 (µij) µij = µ0 + ai + bj + γij γ ∼ Φ(·)
  • 9. Overview ...1 Multilevel Models ...2 Decision-eoretic Model Assessment Framework ...3 Data and Model ...4 Results . . . . . .
  • 10. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results True model, Pseudo-true model and Actual Belief model We assume there is a true underlying model pt(·), from which the observations (both available and future observations) come from. While acknowledging the fact that the true distribution is never accessible, some researchers propose basing the discussion on a rich enough Actual Belief Model), which supposedly fully re ects the uncertainty of future data. (Bernardo and Smith 1994)
  • 11. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results M-closed, M-completed and M-open views In M-closed view, it is assumed that the true model is included in a enumerable collection of models, and the Actual Belief Model is the Bayesian Model Averaging predictive distribution. In M-completed view, the Actual Belief Model p(˜y|D, M) is considered to be the best available description of the uncertainty of future data. In M-open view, the correct speci cation of the Actual Belief Model is avoided and the strategy is to generate Monte Carlo samples from it, such as sample re-use methods.
  • 12. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results A Decision-eoretical Framework We de ne a loss function l(˜y, aM), which is the loss incurred from our inferential action aM, based on a model M, in face of future observation ˜y. en the predictive loss from our inferential action aM is Lp(pt , M, D, l) = Ept(˜y)l(˜y, aM) = ∫ l(˜y, aM)pt (˜y)d˜y It is oen convenient and theoretically desirable to use the whole posterior predictive distribution as aM and the log loss as l(·, ·). Lpred(pt,M,D)=Ept [− log p(˜y|D,M)]=− ∫ pt(˜y) log p(˜y|D,M)d˜y
  • 13. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Decision-eoretic Framework Cont'd For Model Selection task, from a pool of candidate models {Mk : k ∈ K}, we should select the model that minimizes the expected predictive loss. min Mk:k∈K − ∫ pt (˜y) log p(˜y|D, M)d˜y
  • 14. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Decision-eoretic Framework Cont'd For Model Selection task, from a pool of candidate models {Mk : k ∈ K}, we should select the model that minimizes the expected predictive loss. min Mk:k∈K − ∫ pt (˜y) log p(˜y|D, M)d˜y For Model Assessment task of a particular model M, we look at the Kullback-Leibler divergence between the true model and the posterior predictive distribution. We call it the predictive error. Err(pt , M, D) = − ∫ pt (˜y) log p(˜y|D, M)d˜y + ∫ pt (˜y) log pt (˜y)d˜y = KL(p(·|D, M); pt (·))
  • 15. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Estimating Expected Predictive Loss e central obstacle of getting the Expected Predicitve Loss is that we don't know the true distribution pt(·).
  • 16. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Estimating Expected Predictive Loss e central obstacle of getting the Expected Predicitve Loss is that we don't know the true distribution pt(·). A M-closed or M-completed view will substitute the true distribution with a reference distribution.
  • 17. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Estimating Expected Predictive Loss e central obstacle of getting the Expected Predicitve Loss is that we don't know the true distribution pt(·). A M-closed or M-completed view will substitute the true distribution with a reference distribution. From a M-open view, plug in available sample gives us the Training Loss, which has a downward bias, since we used the sample twice. Ltraining(M, D) = − 1 n n∑ i=1 log p(yi|D, M)
  • 18. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Estimating Expected Predictive Loss e central obstacle of getting the Expected Predicitve Loss is that we don't know the true distribution pt(·). A M-closed or M-completed view will substitute the true distribution with a reference distribution. From a M-open view, plug in available sample gives us the Training Loss, which has a downward bias, since we used the sample twice. Ltraining(M, D) = − 1 n n∑ i=1 log p(yi|D, M) ere exist two approaches to get an unbiased estimate of Predictive Loss: Bias Correction which leads to various Information Criteria; Held-out Practices which lead to Leave-one-out Cross Validation and k-fold Cross Validation.
  • 19. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Estimation Methods ere is a long list of variants of Information Criteria, AIC/BIC/DIC/TIC/NIC/WAIC et al.
  • 20. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Estimation Methods ere is a long list of variants of Information Criteria, AIC/BIC/DIC/TIC/NIC/WAIC et al. LOO Cross Validation has been shown to be asymptotically equivalent to AIC/WAIC. But the computational burden is huge. e Importance Sampling method introduces new problem of the reliability of the importance weights.
  • 21. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Estimation Methods ere is a long list of variants of Information Criteria, AIC/BIC/DIC/TIC/NIC/WAIC et al. LOO Cross Validation has been shown to be asymptotically equivalent to AIC/WAIC. But the computational burden is huge. e Importance Sampling method introduces new problem of the reliability of the importance weights. We are using the computationally convenient k-fold cross validation, in which the data set is randomly partitioned into k parts, and in each fold, one part is used as the testing set while the rest serve as the training set.
  • 22. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results k-fold Cross Validation en the k-fold Cross Validation estimate of the Predictive Loss is given by LCV(M, D) = − K∑ k=1 ∑ i∈testk log p(yi|Dk , M) = − N∑ i=1 log p(yi|D(i) , M)
  • 23. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results k-fold Cross Validation en the k-fold Cross Validation estimate of the Predictive Loss is given by LCV(M, D) = − K∑ k=1 ∑ i∈testk log p(yi|Dk , M) = − N∑ i=1 log p(yi|D(i) , M) To estimate the Predictive Error, we still need an estimate of the Entropy of the true distribution. We can use the training loss of the saturated model as a surrogate. − ∫ pt(˜y) log pt(˜y)d˜y = − 1 n n∑ i=1 log p(˙yi|D, Msaturated)
  • 24. Overview ...1 Multilevel Models ...2 Decision-eoretic Model Assessment Framework ...3 Data and Model ...4 Results . . . . . .
  • 25. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Data Set Cooperative Congressional Election Survey 2006 N=30,000 71 social and political response outcomes Deeply nested demographic variables, e.g., state, inc, edu, ethn, gender et al.
  • 26. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Data Set Cont'd Figure: A sample of the questions in CCES 2006 survey.
  • 27. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Model Setup For demonstration, we only consider two demographic variables, state and income, together with their interaction. e responses are all yes-no binary outcomes. Complete Pooling πj1j2 = logit−1 ( βstt j1 + βinc j2 ) No Pooling πj1j2 = logit−1 ( βstt j1 + βinc j2 + βstt*inc j1j2 ) Partial Pooling πj1j2 = logit−1 ( βstt j1 + βinc j2 + βstt*inc j1j2 ) βstt*inc j1j2 ∼ Φ(·)
  • 28. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results k-fold Cross Validation Estimate Due to computational constraints, we are using Maximum A Posteriori plug-in estimate instead of full Bayesian estimate. p(˜y|D, M) ≈ p(˜y|ˆπij(D), M)
  • 29. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results k-fold Cross Validation Estimate Due to computational constraints, we are using Maximum A Posteriori plug-in estimate instead of full Bayesian estimate. p(˜y|D, M) ≈ p(˜y|ˆπij(D), M) en under the aforementioned setup, the Cross Validation estimate of the Predictive Loss is LCV(M,D)=− 1 N ∑K k=1 ∑ l∈testk log p(yl|Dk,M) =− 1 N ∑K k=1 ∑ i,j[y testk ij log ˆπij(Dtraink )+(n testk ij −y testk ij ) log(1−ˆπij(Dtraink ))] =− 1 N ∑ i,j ∑K k=1[log ˆπij(Dtraink )y testk ij +log(1−ˆπij(Dtraink ))(n testk ij −y testk ij )] =− 1 N ∑ i,j [ log ˆπij(Dtrain)yij+log(1−ˆπij(Dtrain))(nij−yij) ] =− ∑ i,j nij N [ log ˆπij(Dtrain)˜πij+log(1−ˆπij(Dtrain))(1−˜πij) ]
  • 30. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Calibration of Improvement Let's suppose we only have one cell, with true proportion .4, and the good model gives a posterior estimate of log proportion at roughly log(0.41), and the lesser model gives a estimate of log(0.44) or log(0.38). en the Predictive Loss under the good model is −[.4 ∗ log(.41) + .6 ∗ log(.59)] = 0.67322, and under the two lesser models is −[.4 ∗ log(.44) + .6 ∗ log(.56)] = 0.67386 and −[.4 ∗ log(.38) + .6 ∗ log(.62)] = 0.67628. We can see the improvement of the Predictive Loss is between 0.0006 to 0.003. Also, the lower bound is given by −[.4 ∗ log(.4) + .6 ∗ log(.6)] = 0.67301, so the Predictive Error of the good model is about 0.0002.
  • 31. Overview ...1 Multilevel Models ...2 Decision-eoretic Model Assessment Framework ...3 Data and Model ...4 Results . . . . . .
  • 32. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Cross Validation Results on All Outcomes Responses (ordered by the lower bound) EstimatedPredictiveError 0.01 0.02 0.03 0.04 0.05 10 20 30 40 50 60 70 models complete pooling partial pooling no pooling Figure: Measure of t (Estimated Predictive Error) for all response outcomes in CCES 2006 survey data. Responses are ordered by the lower bound (training loss of the saturated model). No Pooling model gives very bad t, while Predictive Error of Partial Pooling is dominated by Complete Pooling, but the differences seem small.
  • 33. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Compare Partial Pooling and Complete Pooling In the previous gure, apparently No Pooling is doing very badly, but the differences between Partial Pooling and Complete Pooling seem small. We need to further calibrate them.
  • 34. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Compare Partial Pooling and Complete Pooling In the previous gure, apparently No Pooling is doing very badly, but the differences between Partial Pooling and Complete Pooling seem small. We need to further calibrate them. e summary of the differences between Partial Pooling and Complete Pooling for all the outcomes is Min. 1st Qu. Median Mean 3rd Qu. Max. -0.0003405 0.0001821 0.0003827 0.0006041 0.0005630 0.0053770
  • 35. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Compare Partial Pooling and Complete Pooling In the previous gure, apparently No Pooling is doing very badly, but the differences between Partial Pooling and Complete Pooling seem small. We need to further calibrate them. e summary of the differences between Partial Pooling and Complete Pooling for all the outcomes is Min. 1st Qu. Median Mean 3rd Qu. Max. -0.0003405 0.0001821 0.0003827 0.0006041 0.0005630 0.0053770 We can see that the improvement in terms of the Predictive Loss indeed corresponds to some meaningful improvement in prediction accuracy.
  • 36. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Simulations Based on Real Data We want to explore how the structure of the multilevel models affects the dynamics of the performance of different models. Speci cally, we are interested in total sample size and how balanced the cells are in terms of cell size.
  • 37. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Simulations Based on Real Data We want to explore how the structure of the multilevel models affects the dynamics of the performance of different models. Speci cally, we are interested in total sample size and how balanced the cells are in terms of cell size. We generated simulated data sets based on the real data set, i.e., we use the estimated from the Multilevel model t of the real data sets and enlarge the total sample size by 2, 3 and 4 times, either keeping the original relative proportions (highly unequal) of different cells or making the proportions roughly equal.
  • 38. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Simulation Results: Total Sample Size Responses (ordered by the lower bound) EstimatedPredictiveError 0.002 0.003 0.004 0.005 10 20 30 40 50 60 70 models complete pooling partial pooling no pooling Responses (ordered by the lower bound) EstimatedPredictiveError 0.0020 0.0025 0.0030 0.0035 0.0040 0.0045 10 20 30 40 50 60 70 models complete pooling partial pooling no pooling Responses (ordered by the lower bound) EstimatedPredictiveError 0.002 0.003 0.004 0.005 0.006 10 20 30 40 50 60 70 models complete pooling partial pooling no pooling Figure: Estimated Predictive Error of all response outcomes for ``augmented'' data sets.
  • 39. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Simulation Results: Total Sample Size on House Rep Vote sample size EstimatedPredictiveError 0.002 0.004 0.006 0.008 0.010 0.012 0.014 50000 100000 150000 200000 models complete pooling partial pooling no pooling Figure: Predictive Error of the three models as sample size grows. e outcome under consideration is the Republican vote in the House election.
  • 40. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Simulation Results: Balancedness of the Structure Responses (ordered by the lower bound) EstimatedPredictiveError 0.010 0.015 0.020 0.025 0.030 10 20 30 40 50 60 70 models complete pooling partial pooling no pooling Figure: Measure of t (Predictive Error) for all responses, ordered by lower bound. e data set is simulated from real data set, and has the same sample size in total as the real data set, but keeping all demographic-geographic cells balanced.
  • 41. . . . . . . Multilevel Models Decision-eoretic Model Assessment Framework Data and Model Results Conclusions Cross-validation is not a very sensitive instrument in comparing multilevel models. Careful calibrations are needed for better understanding of the results. We also explored how different aspects of the data set structure affect the margin of improvement.