Sparse vs.

Ensemble Approaches
to Supervised Learning
Greg Grudic
Modified by Longin Jan Latecki
Some slides are by Piyush Rai

Intro AI Ensembles 1
 Goal of Supervised Learning?
• Minimize the probability of model
prediction errors on future data

• Two Competing Methodologies

– Build one really good model
• Traditional approach
– Build many models and average the results
• Ensemble learning (more recent)

Intro AI Ensembles 2
 The Single Model Philosophy
• Motivation: Occam’s Razor
– “one should not increase, beyond what is
necessary, the number of entities required to
explain anything”
• Infinitely many models can explain any
given dataset
– Might as well pick the smallest one…

Intro AI Ensembles 3
 Which Model is Smaller?

• In this case

• It’s not always easy to define small!

Intro AI Ensembles 4
 Exact Occam’s Razor Models
• Exact approaches find optimal solutions
• Examples:
– Support Vector Machines
• Find a model structure that uses the smallest
percentage of training data (to explain the rest of it).
– Bayesian approaches
• Minimum description length

Intro AI Ensembles 5
 How Do Support Vector Machines
Define Small?
Minimize the number
of Support Vectors!

Maximized
Margin

Intro AI Ensembles 6
 Approximate Occam’s Razor
Models
• Approximate solutions use a greedy search
approach which is not optimal
• Examples
– Kernel Projection Pursuit algorithms
• Find a minimal set of kernel projections
– Relevance Vector Machines
• Approximate Bayesian approach
– Sparse Minimax Probability Machine Classification
• Find a minimum set of kernels and features

Intro AI Ensembles 7
 Other Single Models: Not
Necessarily Motivated by Occam’s
Razor
• Minimax Probability Machine (MPM)
• Trees
– Greedy approach to sparseness
• Neural Networks
• Nearest Neighbor
• Basis Function Models
– e.g. Kernel Ridge Regression
Intro AI Ensembles 8
 Ensemble Philosophy
• Build many models and combine them
• Only through averaging do we get at the
truth!
• It’s too hard (impossible?) to build a single
model that works best
• Two types of approaches:
– Models that don’t use randomness
– Models that incorporate randomness

Intro AI Ensembles 9
 Ensemble Approaches
• Bagging
– Bootstrap aggregating

• Boosting

• Random Forests
– Bagging reborn

Intro AI Ensembles 10
 Bagging
• Main Assumption:
– Combining many unstable predictors to produce a
ensemble (stable) predictor.
– Unstable Predictor: small changes in training data
produce large changes in the model.
• e.g. Neural Nets, trees
• Stable: SVM (sometimes), Nearest Neighbor.
• Hypothesis Space
– Variable size (nonparametric):
• Can model any function if you use an appropriate predictor
(e.g. trees)

Intro AI Ensembles 11
 The Bagging Algorithm

Given data:

For
• Obtain bootstrap sample from the
training data
• Build a model from bootstrap data

Intro AI Ensembles 12
 The Bagging Model
• Regression

• Classification:
– Vote over classifier outputs

Intro AI Ensembles 13
 Bagging Details
• Bootstrap sample of N instances is obtained
by drawing N examples at random, with
replacement.
• On average each bootstrap sample has
63% of instances
– Encourages predictors to have
uncorrelated errors
• This is why it works
Intro AI Ensembles 14
 Bagging Details 2
• Usually set
– Or use validation data to pick
• The models need to be unstable
– Usually full length (or slightly pruned) decision
trees.

Intro AI Ensembles 15
 Boosting
– Main Assumption:
• Combining many weak predictors (e.g. tree stumps
or 1-R predictors) to produce an ensemble predictor
• The weak predictors or classifiers need to be stable
– Hypothesis Space
• Variable size (nonparametric):
– Can model any function if you use an appropriate
predictor (e.g. trees)

Intro AI Ensembles 16
 Commonly Used Weak Predictor
(or classifier)
A Decision Tree Stump (1-R)

Intro AI Ensembles 17
 Boosting

Each classifier is
trained from a weighted
Sample of the training
Data

Intro AI Ensembles 18
 Boosting (Continued)
• Each predictor is created by using a biased
sample of the training data
– Instances (training examples) with high error
are weighted higher than those with lower error
• Difficult instances get more attention
– This is the motivation behind boosting

Intro AI Ensembles 19
 Background Notation
• The function is defined as:

• The function is the natural logarithm

Intro AI Ensembles 20
 The AdaBoost Algorithm
(Freund and Schapire, 1996)
Given data:

1. Initialize weights
2. For
a) Fit classifier to data using weights
b) Compute

c) Compute
d) Set
Intro AI Ensembles 21
 The AdaBoost Model

AdaBoost is NOT used for Regression!

Intro AI Ensembles 22
 The Updates in Boosting

Intro AI Ensembles 23
 Boosting Characteristics
Simulated data: test error
rate for boosting with
stumps, as a function of
the number of iterations.
Also shown are the test
error rate for a single
stump, and a 400 node
tree.

Intro AI Ensembles 24
 Loss Functions for
•Misclassification

•Exponential (Boosting)

•Binomial Deviance
(Cross Entropy)

•Squared Error

Incorrect Classification Correct Classification

•Support Vectors

Intro AI Ensembles 25
Intro AI Ensembles 26
Intro AI Ensembles 27
Intro AI Ensembles 28
Intro AI Ensembles 29
Intro AI Ensembles 30
Intro AI Ensembles 31
 Other Variations of Boosting
• Gradient Boosting
– Can use any cost function
• Stochastic (Gradient) Boosting
– Bootstrap Sample: Uniform random sampling
(with replacement)
– Often outperforms the non-random version

Intro AI Ensembles 32
 Gradient Boosting

Intro AI Ensembles 33
 Boosting Summary
• Good points
– Fast learning
– Capable of learning any function (given appropriate weak learner)
– Feature weighting
– Very little parameter tuning
• Bad points
– Can overfit data
– Only for binary classification
• Learning parameters (picked via cross validation)
– Size of tree
– When to stop
• Software
– http://www-stat.stanford.edu/~jhf/R-MART.html

Intro AI Ensembles 34