Download EvaluAtion

Model Evaluation
CRISP-DM
CRISP-DM Phases
• Business Understanding
– Initial phase
– Focuses on:
• Understanding the project objectives and requirements from a business
perspective
• Converting this knowledge into a data mining problem definition, and a
preliminary plan designed to achieve the objectives
• Data Understanding
– Starts with an initial data collection
– Proceeds with activities aimed at:
•
•
•
•
Getting familiar with the data
Identifying data quality problems
Discovering first insights into the data
Detecting interesting subsets to form hypotheses for hidden information
CRISP-DM Phases
• Data Preparation
– Covers all activities to construct the final dataset (data that will be fed
into the modeling tool(s)) from the initial raw data
– Data preparation tasks are likely to be performed multiple times, and
not in any prescribed order
– Tasks include table, record, and attribute selection, as well as
transformation and cleaning of data for modeling tools
• Modeling
– Various modeling techniques are selected and applied, and their
parameters are calibrated to optimal values
– Typically, there are several techniques for the same data mining
problem type
– Some techniques have specific requirements on the form of data,
therefore, stepping back to the data preparation phase is often needed
CRISP-DM Phases
• Evaluation
– At this stage, a model (or models) that appears to have
high quality, from a data analysis perspective, has been
built
– Before proceeding to final deployment of the model, it is
important to more thoroughly evaluate the model, and
review the steps executed to construct the model, to be
certain it properly achieves the business objectives
– A key objective is to determine if there is some important
business issue that has not been sufficiently considered
– At the end of this phase, a decision on the use of the data
mining results should be reached
CRISP-DM Phases
• Deployment
– Creation of the model is generally not the end of the project
– Even if the purpose of the model is to increase knowledge of the data,
the knowledge gained will need to be organized and presented in a
way that the customer can use it
– Depending on the requirements, the deployment phase can be as
simple as generating a report or as complex as implementing a
repeatable data mining process
– In many cases it will be the customer, not the data analyst, who will
carry out the deployment steps
– However, even if the analyst will not carry out the deployment effort it
is important for the customer to understand up front what actions will
need to be carried out in order to actually make use of the created
models
Evaluating Classification Systems
• Two issues
– What evaluation measure should we use?
– How do we ensure reliability of our model?
How do we ensure reliability of our model?
EVALUATION
How do we ensure reliability?
• Heavily dependent on training
Data Partitioning
• Randomly partition data into training and test set
• Training set – data used to train/build the model.
– Estimate parameters (e.g., for a linear regression), build decision tree, build
artificial network, etc.
• Test set – a set of examples not used for model induction. The model’s
performance is evaluated on unseen data. Aka out-of-sample data.
• Generalization Error: Model error on the test data.
Set of training examples
Set of test
examples
Complexity and Generalization
Score
Function
e.g., squared
error
Optimal model
complexity
Stest(q)
Strain(q)
Complexity = degrees
of freedom in the model
(e.g., number of variables)
Holding out data
• The holdout method reserves a certain amount for
testing and uses the remainder for training
– Usually: one third for testing, the rest for training
• For “unbalanced” datasets, random samples might not
be representative
– Few or none instances of some classes
• Stratified sample:
– Make sure that each class is represented with approximately
equal proportions in both subsets
12
Repeated holdout method
• Holdout estimate can be made more reliable by
repeating the process with different subsamples
– In each iteration, a certain proportion is randomly
selected for training (possibly with stratification)
– The error rates on the different iterations are
averaged to yield an overall error rate
• This is called the repeated holdout method
13
Cross-validation
• Most popular and effective type of repeated holdout is
cross-validation
• Cross-validation avoids overlapping test sets
– First step: data is split into k subsets of equal size
– Second step: each subset in turn is used for testing and the
remainder for training
• This is called k-fold cross-validation
• Often the subsets are stratified before the crossvalidation is performed
14
Cross-validation example:
15
15
More on cross-validation
• Standard data-mining method for evaluation: stratified
ten-fold cross-validation
• Why ten? Extensive experiments have shown that this
is the best choice to get an accurate estimate
• Stratification reduces the estimate’s variance
• Even better: repeated stratified cross-validation
– E.g. ten-fold cross-validation is repeated ten times and results
are averaged (reduces the sampling variance)
• Error estimate is the mean across all repetitions
16
Leave-One-Out cross-validation
•
Leave-One-Out:
a particular form of cross-validation:
–
–
•
•
•
17
Set number of folds to number of training instances
I.e., for n training instances, build classifier n times
Makes best use of the data
Involves no random subsampling
Computationally expensive, but good performance
Leave-One-Out-CV and stratification
•
Disadvantage of Leave-One-Out-CV: stratification is not
possible
–
•
Extreme example: random dataset split equally into two classes
–
–
–
18
It guarantees a non-stratified sample because there is only one instance
in the test set!
Best model predicts majority class
50% accuracy on fresh data
Leave-One-Out-CV estimate is 100% error!
Three way data splits
• One problem with CV is since data is being used jointly
to fit model and estimate error, the error could be
biased downward.
• If the goal is a real estimate of error (as opposed to
which model is best), you may want a three way split:
– Training set: examples used for learning
– Validation set: used to tune parameters
– Test set: never used in the model fitting process, used at the
end for unbiased estimate of hold out error
The Bootstrap
• The Statistician Brad Efron proposed a very
simple and clever idea for mechanically
estimating confidence intervals:
The Bootstrap
• The idea is to take multiple resamples of your
original dataset.
• Compute the statistic of interest on each
resample
• you thereby estimate the distribution of this statistic!
Sampling with Replacement
• Draw a data point at random from the data set.
• Then throw it back in
• Draw a second data point.
• Then throw it back in…
• Keep going until we’ve got 1000 data points.
• You might call this a “pseudo” data set.
• This is not merely re-sorting the data.
• Some of the original data points will appear more than once;
others won’t appear at all.
Sampling with Replacement
• In fact, there is a chance of
(1-1/1000)1000 ≈ 1/e ≈ .368
that any one of the original data points won’t appear at all
if we sample with replacement 1000 times.
 any data point is included with Prob ≈ .632
• Intuitively, we treat the original sample as the “true
population in the sky”.
• Each resample simulates the process of taking a sample
from the “true” distribution.
Bootstrapping & Validation
• This is interesting in its own right.
• But bootstrapping also relates back to model validation.
• Along the lines of cross-validation.
• You can fit models on bootstrap resamples of your
data.
• For each resample, test the model on the ≈ .368 of the
data not in your resample.
• Will be biased, but corrections are available.
• Get a spectrum of ROC curves.
Closing Thoughts
• The “cross-validation” approach has several nice
features:
– Relies on the data, not likelihood theory, etc.
– Comports nicely with the lift curve concept.
– Allows model validation that has both business & statistical
meaning.
– Is generic  can be used to compare models generated from
competing techniques…
– … or even pre-existing models
– Can be performed on different sub-segments of the data
– Is very intuitive, easily grasped.
Closing Thoughts
• Bootstrapping has a family resemblance to cross-
validation:
– Use the data to estimate features of a statistic or a model that we
previously relied on statistical theory to give us.
– Classic examples of the “data mining” (in the non-pejorative sense of
the term!) mindset:
• Leverage modern computers to “do it yourself” rather than look up a
formula in a book!
• Generic tools that can be used creatively.
– Can be used to estimate model bias & variance.
– Can be used to estimate (simulate) distributional characteristics of
very difficult statistics.
– Ideal for many actuarial applications.
What evaluation measure should we use?
METRICS
Evaluation of Classification
actual
outcome
1
0
– Not always the best choice
• Assume 1% fraud,
• model predicts no fraud
• What is the accuracy?
1
a
b
0
c
d
predicted
outcome
Accuracy = (a+d) / (a+b+c+d)
Actual Class
Predicted Class
Fraud
No Fraud
Fraud
0
0
No Fraud
10
990
Evaluation of Classification
Other options:
– recall or sensitivity (how many of those that are really positive did you
predict?):
• a/(a+c)
– precision (how many of those predicted positive really are?)
• a/(a+b)
actual
outcome
Precision and recall are always in tension
1
0
1
a
b
0
c
d
– Increasing one tends to decrease another
predicted
outcome
Evaluation of Classification
Yet another option:
– recall or sensitivity (how many of the positives did you get right?):
• a/(a+c)
– Specificity (how many of the negatives did you get right?)
• d/(b+d)
actual
outcome
Sensitivity and specificity have the same tension
Different fields use different metrics
1
0
1
a
b
0
c
d
predicted
outcome
Evaluation for a Thresholded
Response
• Many classification models
output probabilities
• These probabilities get
thresholded to make a
prediction.
• Classification accuracy
depends on the threshold –
good models give low
probabilities to Y=0 and high
probabilities to Y=1.
predicted probabilities
Suppose we use a cutoff
of 0.5…
actual outcome
1
1
predicted
outcome
0
8
3
0
9
0
Test Data
Suppose we use a cutoff
of 0.5…
actual outcome
1
0
sensitivity:
1
predicted
outcome
8
= 100%
3
specificity:
0
0
8
8+0
9
9+3
= 75%
9
we want both of these to be high
Suppose we use a cutoff
of 0.8…
actual outcome
1
0
sensitivity:
1
predicted
outcome
6
= 75%
10
10+2
= 83%
2
specificity:
0
2
6
6+2
10
•
Note there are 20 possible thresholds
•
Plotting all values of sensitivity vs. specificity gives a sense
of model performance by seeing the tradeoff with
different thresholds
•
Note if threshold = minimum
actual outcome
c=d=0 so sens=1; spec=0
•
0
a
b
c
d
If threshold = maximum
a=b=0 so sens=0; spec=1
•
1
1
If model is perfect
sens=1; spec=1
0
ROC curve plots sensitivity vs.
(1-specificity) – also known as
false positive rate
Always goes from (0,0) to (1,1)
The more area in the upper left,
the better
Random model is on the
diagonal
“Area under the curve” (AUC)
is a common measure of
predictive performance
ROC CURVES
Receiver Operating Characteristic curve
• ROC curves were developed in the 1950's as a by-product of research into
making sense of radio signals contaminated by noise. More recently it's
become clear that they are remarkably useful in decision-making.
• They are a performance graphing method.
• True positive and False positive fractions are plotted as we move the dividing
threshold. They look like:
• ROC graphs are two-dimensional
graphs in which TP rate is plotted on
the Y axis and FP rate is plotted on the
X axis.
• An ROC graph depicts relative tradeoffs between benefits (true positives)
and costs (false positives).
• Figure shows an ROC graph with five
classifiers labeled A through E.
• A discrete classier is one that outputs
only a class label.
• Each discrete classier produces an (fp
rate, tp rate) pair corresponding to a
single point in ROC space.
• Classifiers in figure are all discrete
classifiers.
ROC Space
Several Points in ROC Space
• Lower left point (0, 0) represents the
strategy of never issuing a positive
classification;
– such a classier commits no false positive
errors but also gains no true positives.
• Upper right corner (1, 1) represents the
opposite strategy, of unconditionally
issuing positive classifications.
• Point (0, 1) represents perfect
classification.
– D's performance is perfect as shown.
• Informally, one point in ROC space is
better than another if it is to the
northwest of the first
– tp rate is higher, fp rate is lower, or both.
Specific Example
Pts without
the disease
Pts with
disease
Test Result
Threshold
Call these patients “negative”
Call these patients “positive”
Test Result
Some definitions ...
Call these patients “negative”
Call these patients “positive”
True Positives
Test Result
without the disease
with the disease
Call these patients “negative”
Call these patients “positive”
Test Result
without the disease
with the disease
False
Positives
Call these patients “negative”
Call these patients “positive”
True
negatives
Test Result
without the disease
with the disease
Call these patients “negative”
Call these patients “positive”
False
negatives
Test Result
without the disease
with the disease
Moving the Threshold: right
‘‘’’
‘‘+
’’
Test Result
without the disease
with the disease
Moving the Threshold: left
‘‘’’
‘‘+
’’
Test Result
without the disease
with the disease
ROC curve
True Positive Rate
(sensitivity)
100%
0%
0%
False Positive Rate
(1-specificity)
100%
ROC curve comparison
A poor test:
A good test:
100%
True Positive Rate
True Positive Rate
100%
0%
0%
0%
100%
False Positive Rate
0
%
100%
False Positive Rate
ROC curve extremes
Best Test:
Worst test:
100%
True Positive Rate
True Positive Rate
100%
0
%
0
%
0
%
False Positive Rate
100
%
The distributions
don’t overlap at
all
0
%
False Positive
Rate
100
%
The distributions
overlap completely
How to Construct ROC Curve for one
Classifier
•
•
•
•
Ppos
0.99
0.98
0.7
0.6
0.43
Sort the instances according to their Ppos.
Move a threshold on the sorted instances.
For each threshold define a classifier with confusion matrix.
Plot the TPr and FPr rates of the classfiers.
True Class
pos
pos
neg
pos
neg
Predicted
True
pos
neg
pos
2
1
neg
1
1
Creating an ROC Curve
• A classifier produces a single ROC point.
• If the classifier has a “sensitivity” parameter,
varying it produces a series of ROC points
(confusion matrices).
• Alternatively, if the classifier is produced by a
learning algorithm, a series of ROC points can
be generated by varying the class ratio in the
training set.
ROC for one Classifier
Good separation between the classes, convex curve.
ROC for one Classifier
Reasonable separation between the classes, mostly convex.
ROC for one Classifier
Fairly poor separation between the classes, mostly convex.
ROC for one Classifier
Poor separation between the classes, large and small concavities.
ROC for one Classifier
Random performance.
The AUC Metric
• The area under ROC curve (AUC) assesses the ranking in terms
of separation of the classes.
• AUC estimates that randomly chosen positive instance will be
ranked before randomly chosen negative instances.
Comparing Models
• Highest AUC wins
• But pay attention to
‘Occam’s Razor’
– ‘the best theory is the
smallest one that describes
all the facts’
– Also known as the
‘parsimony principle’
– If two models are similar,
pick the simpler one