Classification: Basic Concepts,

Decision Trees, and Model

Evaluation

Lecture Notes for Chapter 4

Introduction to Data Mining

by
Tan, Steinbach, Kumar

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 1

 Classification: Definition

 Given a collection of records (training set )

– Each record contains a set of attributes, one of the
attributes is the class.
 Find a model for class attribute as a function
of the values of other attributes.
 Goal: previously unseen records should be
assigned a class as accurately as possible.
– A test set is used to determine the accuracy of the
model. Usually, the given data set is divided into
training and test sets, with training set used to build
the model and test set used to validate it.

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 2

 Illustrating Classification Task

Tid Attrib1 Attrib2 Attrib3 Class Learning

No
1 Yes Large 125K
algorithm
2 No Medium 100K No

3 No Small 70K No
4 Yes Medium 120K No
Induction
5 No Large 95K Yes
6 No Medium 60K No

7 Yes Large 220K No Learn

8 No Small 85K Yes Model
9 No Medium 75K No
10 No Small 90K Yes
Model
10

Training Set
Apply
Tid Attrib1 Attrib2 Attrib3 Class Model
11 No Small 55K ?

12 Yes Medium 80K ?

13 Yes Large 110K ? Deduction

14 No Small 95K ?

15 No Large 67K ?
10

Test Set

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 3

 Examples of Classification Task

 Predicting tumor cells as benign or malignant

 Classifying credit card transactions

as legitimate or fraudulent

 Classifying secondary structures of protein

as alpha-helix, beta-sheet, or random
coil

 Categorizing news stories as finance,

weather, entertainment, sports, etc
© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 4
 Classification Techniques

 Decision Tree based Methods

 Rule-based Methods
 Memory based reasoning
 Neural Networks
 Naïve Bayes and Bayesian Belief Networks
 Support Vector Machines

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 5

 Example of a Decision Tree

cal cal us
i i o
or or nu
teg
teg
nti
ass
ca ca co cl
Tid Refund Marital Taxable
Splitting Attributes
Status Income Cheat

1 Yes Single 125K No

2 No Married 100K No Refund
3 No Single 70K No
Yes No
4 Yes Married 120K No NO MarSt
5 No Divorced 95K Yes Married
Single, Divorced
6 No Married 60K No
7 Yes Divorced 220K No TaxInc NO
8 No Single 85K Yes < 80K > 80K
9 No Married 75K No
NO YES
10 No Single 90K Yes
10

Training Data Model: Decision Tree

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 6

 Another Example of Decision Tree

cal cal us
i i o
or or nu
teg
teg
nti
ass Single,
ca ca co cl MarSt
Married Divorced
Tid Refund Marital Taxable
Status Income Cheat
NO Refund
1 Yes Single 125K No
Yes No
2 No Married 100K No
3 No Single 70K No NO TaxInc
4 Yes Married 120K No < 80K > 80K
5 No Divorced 95K Yes
NO YES
6 No Married 60K No
7 Yes Divorced 220K No
8 No Single 85K Yes
9 No Married 75K No There could be more than one tree that
10 No Single 90K Yes fits the same data!
10

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 7

 Decision Tree Classification Task

Tid Attrib1 Attrib2 Attrib3 Class

Tree
1 Yes Large 125K No Induction
2 No Medium 100K No algorithm
3 No Small 70K No
4 Yes Medium 120K No
Induction
5 No Large 95K Yes
6 No Medium 60K No

7 Yes Large 220K No Learn

8 No Small 85K Yes Model
9 No Medium 75K No
10 No Small 90K Yes
Model
10

Training Set
Apply Decision
Tid Attrib1 Attrib2 Attrib3 Class
Model Tree
11 No Small 55K ?

12 Yes Medium 80K ?

13 Yes Large 110K ?

Deduction
14 No Small 95K ?

15 No Large 67K ?
10

Test Set

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 8

 Apply Model to Test Data

Test Data
Start from the root of tree. Refund Marital Taxable
Status Income Cheat

No Married 80K ?
Refund 10

Yes No

NO MarSt
Single, Divorced Married

TaxInc NO
< 80K > 80K

NO YES

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 9

 Apply Model to Test Data

Test Data
Refund Marital Taxable
Status Income Cheat

No Married 80K ?
Refund 10

Yes No

NO MarSt
Single, Divorced Married

TaxInc NO
< 80K > 80K

NO YES

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 10

 Apply Model to Test Data

Test Data
Refund Marital Taxable
Status Income Cheat

No Married 80K ?
Refund 10

Yes No

NO MarSt
Single, Divorced Married

TaxInc NO
< 80K > 80K

NO YES

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 11

 Apply Model to Test Data

Test Data
Refund Marital Taxable
Status Income Cheat

No Married 80K ?
Refund 10

Yes No

NO MarSt
Single, Divorced Married

TaxInc NO
< 80K > 80K

NO YES

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 12

 Apply Model to Test Data

Test Data
Refund Marital Taxable
Status Income Cheat

No Married 80K ?
Refund 10

Yes No

NO MarSt
Single, Divorced Married

TaxInc NO
< 80K > 80K

NO YES

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 13

 Apply Model to Test Data

Test Data
Refund Marital Taxable
Status Income Cheat

No Married 80K ?
Refund 10

Yes No

NO MarSt
Single, Divorced Married Assign Cheat to “No”

TaxInc NO
< 80K > 80K

NO YES

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 14

 Decision Tree Classification Task

Tid Attrib1 Attrib2 Attrib3 Class

Tree
1 Yes Large 125K No Induction
2 No Medium 100K No algorithm
3 No Small 70K No

4 Yes Medium 120K No

Induction
5 No Large 95K Yes
6 No Medium 60K No

7 Yes Large 220K No Learn

8 No Small 85K Yes Model
9 No Medium 75K No
10 No Small 90K Yes
Model
10

Training Set
Apply Decision
Tid Attrib1 Attrib2 Attrib3 Class
Model Tree
11 No Small 55K ?

12 Yes Medium 80K ?

13 Yes Large 110K ?

Deduction
14 No Small 95K ?

15 No Large 67K ?
10

Test Set

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 15

 Decision Tree Induction

 Many Algorithms:
– Hunt’s Algorithm (one of the earliest)
– CART
– ID3, C4.5
– SLIQ,SPRINT

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 16

 General Structure of Hunt’s
Algorithm
Tid Refund Marital Taxable
 Let Dt be the set of training records Status Income Cheat

that reach a node t 1 Yes Single 125K No

 General Procedure: 2 No Married 100K No
3 No Single 70K No
– If Dt contains records that 4 Yes Married 120K No
belong the same class yt, then t 5 No Divorced 95K Yes

is a leaf node labeled as yt 6 No Married 60K No

7 Yes Divorced 220K No
– If Dt is an empty set, then t is a 8 No Single 85K Yes
leaf node labeled by the default 9 No Married 75K No
class, yd 10
10 No Single 90K Yes

– If Dt contains records that Dt

belong to more than one class,
use an attribute test to split the
data into smaller subsets. ?
Recursively apply the
procedure to each subset.

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 17

 Hunt’s Algorithm

Refund
Don’t
Yes No
Cheat
Don’t Don’t
Cheat Cheat

Refund Refund
Yes No Yes No

Don’t Don’t Marital

Marital Cheat
Cheat Status Status
Single, Single,
Married Married
Divorced Divorced

Don’t Taxable Don’t

Cheat Cheat
Cheat Income
< 80K >= 80K

Don’t Cheat
Cheat
© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 18
 Tree Induction

 Greedy strategy.
– Split the records based on an attribute test
that optimizes certain criterion.

 Issues
– Determine how to split the records
How to specify the attribute test condition?
How to determine the best split?

– Determine when to stop splitting

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 19

 Tree Induction

 Greedy strategy.
– Split the records based on an attribute test
that optimizes certain criterion.

 Issues
– Determine how to split the records
How to specify the attribute test condition?
How to determine the best split?

– Determine when to stop splitting

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 20

 How to Specify Test Condition?

 Depends on attribute types

– Nominal
– Ordinal
– Continuous

 Depends on number of ways to split

– 2-way split
– Multi-way split

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 21

 Splitting Based on Nominal
Attributes

 Multi-way split: Use as many partitions as distinct

values.
CarType
Family Luxury
Sports

 Binary split: Divides values into two subsets.

Need to find optimal partitioning.
CarType CarType
{Sports, OR {Family,
Luxury} {Family} Luxury} {Sports}

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 22

 Splitting Based on Ordinal
Attributes

 Multi-way split: Use as many partitions as distinct

values.
Size
Small Large
Medium

 Binary split: Divides values into two subsets.

Need to find optimal partitioning.
Size Size
{Small,
{Large}
OR {Medium,
{Small}
Medium} Large}

Size
{Small,
 What about this split? Large} {Medium}

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 23

 Splitting Based on Continuous
Attributes

 Different ways of handling

– Discretization to form an ordinal categorical
attribute
 Static – discretize once at the beginning
 Dynamic – ranges can be found by equal interval

bucketing, equal frequency

bucketing
(percentiles), or clustering.

– Binary Decision: (A < v) or (A  v)

 consider all possible splits and finds the best cut
 can be more compute intensive
© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 24
 Splitting Based on Continuous
Attributes

Taxable Taxable
Income Income?
> 80K?
< 10K > 80K
Yes No

[10K,25K) [25K,50K) [50K,80K)

(i) Binary split (ii) Multi-way split

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 25

 Tree Induction

 Greedy strategy.
– Split the records based on an attribute test
that optimizes certain criterion.

 Issues
– Determine how to split the records
How to specify the attribute test condition?
How to determine the best split?

– Determine when to stop splitting

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 26

 How to determine the Best Split

Before Splitting: 10 records of class 0,

10 records of class 1

Own Car Student

Car? Type? ID?

Yes No Family Luxury c1 c20

c10 c11
Sports
C0: 6 C0: 4 C0: 1 C0: 8 C0: 1 C0: 1 ... C0: 1 C0: 0 ... C0: 0
C1: 4 C1: 6 C1: 3 C1: 0 C1: 7 C1: 0 C1: 0 C1: 1 C1: 1

Which test condition is the best?

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 27

 How to determine the Best Split

 Greedy approach:
– Nodes with homogeneous class distribution
are preferred
 Need a measure of node impurity:

C0: 5 C0: 9
C1: 5 C1: 1

Non-homogeneous, Homogeneous,
High degree of impurity Low degree of impurity

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 28

 Measures of Node Impurity

 Gini Index

 Entropy

 Misclassification error

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 29

 How to Find the Best Split
Before Splitting: C0 N00
M0
C1 N01

A? B?
Yes No Yes No

Node N1 Node N2 Node N3 Node N4

C0 N10 C0 N20 C0 N30 C0 N40

C1 N11 C1 N21 C1 N31 C1 N41

M1 M2 M3 M4

M12 M34
Gain = M0 – M12 vs M0 – M34
© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 30
 Measure of Impurity: GINI

 Gini Index for a given node t :

GINI (t ) 1   [
j
p ( j | t )] 2

(NOTE: p( j | t) is the relative frequency of class j at node t).

– Maximum (1 - 1/nc) when records are equally

distributed among all classes, implying least
interesting information
– Minimum (0.0) when all records belong to one class,
implying most interesting information
C1 0 C1 1 C1 2 C1 3
C2 6 C2 5 C2 4 C2 3
Gini=0.000 Gini=0.278 Gini=0.444 Gini=0.500

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 31

 Examples for computing GINI

GINI (t ) 1   [
j
p ( j | t )] 2

C1 0 P(C1) = 0/6 = 0 P(C2) = 6/6 = 1

C2 6 Gini = 1 – P(C1)2 – P(C2)2 = 1 – 0 – 1 = 0

C1 1 P(C1) = 1/6 P(C2) = 5/6

C2 5 Gini = 1 – (1/6)2 – (5/6)2 = 0.278

C1 2 P(C1) = 2/6 P(C2) = 4/6

C2 4 Gini = 1 – (2/6)2 – (4/6)2 = 0.444

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 32

 Splitting Based on GINI

 Used in CART, SLIQ, SPRINT.

 When a node p is split into k partitions (children), the
quality of split is computed as,
k
ni
GINI split  GINI (i )
i 1 n

where, ni = number of records at child i,

n = number of records at node p.

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 33

Binary Attributes: Computing GINI
Index

 Splits into two partitions

 Effect of Weighing partitions:
– Larger and Purer Partitions are sought for.
Parent
B? C1 6
Yes No C2 6
Gini = 0.500
Node N1 Node N2
Gini(N1)
= 1 – (5/6)2 – (2/6)2 N1 N2 Gini(Children)
= 0.194 C1 5 1 = 7/12 * 0.194 +
Gini(N2) C2 2 4 5/12 * 0.528
= 1 – (1/6)2 – (4/6)2 Gini=0.333 = 0.333
= 0.528
© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 34
 Categorical Attributes: Computing Gini
Index

 For each distinct value, gather counts for each class in

the dataset
 Use the count matrix to make decisions

Multi-way split Two-way split

(find best partition of values)

CarType CarType CarType

Family Sports Luxury {Sports, {Family,
{Family} {Sports}
Luxury} Luxury}
C1 1 2 1 3 1 2 2
C1 C1
C2 4 1 1 C2 2 4 C2 1 5
Gini 0.393 Gini 0.400 Gini 0.419

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 35

 Continuous Attributes: Computing Gini
Index

 Use Binary Decisions based on one

value
 Several Choices for the splitting value
– Number of possible splitting values
= Number of distinct values
 Each splitting value has a count matrix
associated with it
– Class counts in each of the
partitions, A < v and A  v
 Simple method to choose best v
– For each v, scan the database to
gather count matrix and compute Taxable
Income
its Gini index
> 80K?
– Computationally Inefficient!
Repetition of work. Yes No

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 36

 Continuous Attributes: Computing Gini
Index...

 For efficient computation: for each attribute,

– Sort the attribute on values
– Linearly scan these values, each time updating the count matrix and
computing gini index
– Choose the split position that has the least gini index

Cheat No No No Yes Yes Yes No No No No

Taxable Income
60 70 75 85 90 95 100 120 125 220
Sorted Values
55 65 72 80 87 92 97 110 122 172 230
Split Positions
<= > <= > <= > <= > <= > <= > <= > <= > <= > <= > <= >
Yes 0 3 0 3 0 3 0 3 1 2 2 1 3 0 3 0 3 0 3 0 3 0

No 0 7 1 6 2 5 3 4 3 4 3 4 3 4 4 3 5 2 6 1 7 0

Gini 0.420 0.400 0.375 0.343 0.417 0.400 0.300 0.343 0.375 0.400 0.420

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 37

 Alternative Splitting Criteria based on
INFO

 Entropy at a given node t:

Entropy (t )   p ( j | t ) log p ( j | t )
j

(NOTE: p( j | t) is the relative frequency of class j at node t).

– Measures homogeneity of a node.
 Maximum (log nc) when records are equally distributed
among all classes implying least information
 Minimum (0.0) when all records belong to one class,

implying most information

– Entropy based computations are similar to the
GINI index computations
© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 38
 Examples for computing Entropy

Entropy (t )   p ( j | t ) log p ( j | t )
j 2

C1 0 P(C1) = 0/6 = 0 P(C2) = 6/6 = 1

C2 6 Entropy = – 0 log 0 – 1 log 1 = – 0 – 0 = 0

C1 1 P(C1) = 1/6 P(C2) = 5/6

C2 5 Entropy = – (1/6) log2 (1/6) – (5/6) log2 (1/6) = 0.65

C1 2 P(C1) = 2/6 P(C2) = 4/6

C2 4 Entropy = – (2/6) log2 (2/6) – (4/6) log2 (4/6) = 0.92

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 39

 Splitting Based on INFO...

 Information Gain:
 n  k
GAIN Entropy ( p )    Entropy (i )  i

 n 
split i 1

Parent Node, p is split into k partitions;

ni is number of records in partition i
– Measures Reduction in Entropy achieved because of
the split. Choose the split that achieves most reduction
(maximizes GAIN)
– Used in ID3 and C4.5
– Disadvantage: Tends to prefer splits that result in large
number of partitions, each being small but pure.

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 40

 Splitting Based on INFO...

 Gain Ratio:
GAIN n n

k
GainRATIO SplitINFO   log
Split i i

SplitINFO
split

n n i 1

Parent Node, p is split into k partitions

ni is the number of records in partition i

– Adjusts Information Gain by the entropy of the

partitioning (SplitINFO). Higher entropy partitioning
(large number of small partitions) is penalized!
– Used in C4.5
– Designed to overcome the disadvantage of Information
Gain
© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 41
 Splitting Criteria based on Classification
Error

 Classification error at a node t :

Error (t ) 1  max P (i | t ) i

 Measures misclassification error made by a node.

 Maximum (1 - 1/nc) when records are equally distributed
among all classes, implying least interesting information
 Minimum (0.0) when all records belong to one class, implying
most interesting information

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 42

 Examples for Computing Error

Error (t ) 1  max P (i | t ) i

C1 0 P(C1) = 0/6 = 0 P(C2) = 6/6 = 1

C2 6 Error = 1 – max (0, 1) = 1 – 1 = 0

C1 1 P(C1) = 1/6 P(C2) = 5/6

C2 5 Error = 1 – max (1/6, 5/6) = 1 – 5/6 = 1/6

C1 2 P(C1) = 2/6 P(C2) = 4/6

C2 4 Error = 1 – max (2/6, 4/6) = 1 – 4/6 = 1/3

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 43

 Comparison among Splitting
Criteria

For a 2-class problem:

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 44

 Misclassification Error vs Gini

A? Parent
C1 7
Yes No
C2 3
Node N1 Node N2 Gini = 0.42

Gini(N1) N1 N2
= 1 – (3/3)2 – (0/3)2 Gini(Children)
C1 3 4 = 3/10 * 0
=0
C2 0 3 + 7/10 * 0.489
Gini(N2) Gini=0.361 = 0.342
= 1 – (4/7)2 – (3/7)2
= 0.489 Gini improves !!

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 45

 Tree Induction

 Greedy strategy.
– Split the records based on an attribute test
that optimizes certain criterion.

 Issues
– Determine how to split the records
How to specify the attribute test condition?
How to determine the best split?

– Determine when to stop splitting

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 46

 Stopping Criteria for Tree
Induction

 Stop expanding a node when all the records

belong to the same class

 Stop expanding a node when all the records have

similar attribute values

 Early termination (to be discussed later)

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 47

 Decision Tree Based Classification

 Advantages:
– Inexpensive to construct
– Extremely fast at classifying unknown records
– Easy to interpret for small-sized trees
– Accuracy is comparable to other classification
techniques for many simple data sets

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 48

 Example: C4.5

 Simple depth-first construction.

 Uses Information Gain
 Sorts Continuous Attributes at each node.
 Needs entire data to fit in memory.
 Unsuitable for Large Datasets.

– Needs out-of-core sorting.

 You can download the software from:

http://www.cse.unsw.edu.au/~quinlan/c4.5r8.tar.gz

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 49

 Practical Issues of Classification

 Underfitting and Overfitting

 Missing Values

 Costs of Classification

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 50

 Underfitting and Overfitting
(Example)

500 circular and 500

triangular data points.

Circular points:
0.5  sqrt(x12+x22)  1

Triangular points:
sqrt(x12+x22) > 0.5 or
sqrt(x12+x22) < 1

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 51

 Underfitting and Overfitting

Overfitting

Underfitting: when model is too simple, both training and test errors are large

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 52

 Overfitting due to Noise

Decision boundary is distorted by noise point

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 53

 Overfitting due to Insufficient
Examples

Lack of data points in the lower half of the diagram makes it difficult
to predict correctly the class labels of that region
- Insufficient number of training records in the region causes the
decision tree to predict the test examples using other training
records that are irrelevant to the classification task
© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 54
 Notes on Overfitting

 Overfitting results in decision trees that are more

complex than necessary

 Training error no longer provides a good estimate

of how well the tree will perform on previously
unseen records

 Need new ways for estimating errors

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 55

 Estimating Generalization Errors
 Re-substitution errors: error on training ( e(t) )
 Generalization errors: error on testing ( e’(t))
 Methods for estimating generalization errors:
– Optimistic approach: e’(t) = e(t)
– Pessimistic approach:
 For each leaf node: e’(t) = (e(t)+0.5)
 Total errors: e’(T) = e(T) + N  0.5 (N: number of leaf nodes)
 For a tree with 30 leaf nodes and 10 errors on training
(out of 1000 instances):
Training error = 10/1000 = 1%
Generalization error = (10 + 300.5)/1000 = 2.5%
– Reduced error pruning (REP):
 uses validation data set to estimate generalization
error

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 56

 Occam’s Razor

 Given two models of similar generalization errors,

one should prefer the simpler model over the
more complex model

 For complex models, there is a greater chance

that it was fitted accidentally by errors in data

 Therefore, one should include model complexity

when evaluating a model

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 57

 Minimum Description Length
(MDL)
A?
X y Yes No
X y
X1 1 0 B? X1 ?
X2 0 B1 B2
X2 ?
X3 0 C? 1
A C1 C2 B X3 ?
X4 1
0 1 X4 ?
… …
Xn
… …
1
Xn ?

 Cost(Model,Data) = Cost(Data|Model) + Cost(Model)

– Cost is the number of bits needed for encoding.
– Search for the least costly model.
 Cost(Data|Model) encodes the misclassification errors.
 Cost(Model) uses node encoding (number of children)
plus splitting condition encoding.
© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 58
 How to Address Overfitting

 Pre-Pruning (Early Stopping Rule)

– Stop the algorithm before it becomes a fully-grown tree
– Typical stopping conditions for a node:
 Stop if all instances belong to the same class
 Stop if all the attribute values are the same
– More restrictive conditions:
 Stop if number of instances is less than some user-specified
threshold
 Stop if class distribution of instances are independent of the
available features (e.g., using  2 test)
 Stop if expanding the current node does not improve impurity
measures (e.g., Gini or information gain).

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 59

 How to Address Overfitting…

 Post-pruning
– Grow decision tree to its entirety
– Trim the nodes of the decision tree in a
bottom-up fashion
– If generalization error improves after trimming,
replace sub-tree by a leaf node.
– Class label of leaf node is determined from
majority class of instances in the sub-tree
– Can use MDL for post-pruning

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 60

 Example of Post-Pruning
Training Error (Before splitting) = 10/30

Class = Yes 20 Pessimistic error = (10 + 0.5)/30 = 10.5/30

Class = No 10 Training Error (After splitting) = 9/30

Error = 10/30 Pessimistic error (After splitting)

= (9 + 4  0.5)/30 = 11/30
PRUNE!
A?

A1 A4
A2 A3

Class = Yes 8 Class = Yes 3 Class = Yes 4 Class = Yes 5

Class = No 4 Class = No 4 Class = No 1 Class = No 1

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 61

 Examples of Post-pruning

– Optimistic error? Case 1:

Don’t prune for both cases

C0: 11 C0: 2
– Pessimistic error? C1: 3 C1: 4

Don’t prune case 1, prune case 2

– Reduced error pruning?

Case 2:

Depends on validation set

C0: 14 C0: 2
C1: 3 C1: 2

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 62

 Handling Missing Attribute Values

 Missing values affect decision tree construction in

three different ways:
– Affects how impurity measures are computed
– Affects how to distribute instance with missing
value to child nodes
– Affects how a test instance with missing value
is classified

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 63

 Computing Impurity Measure

Tid Refund Marital Taxable Before Splitting:

Status Income Class Entropy(Parent)
= -0.3 log(0.3)-(0.7)log(0.7) = 0.8813
1 Yes Single 125K No
2 No Married 100K No Class Class
3 No Single 70K No = Yes = No
Refund=Yes 0 3
4 Yes Married 120K No
Refund=No 2 4
5 No Divorced 95K Yes
Refund=? 1 0
6 No Married 60K No
7 Yes Divorced 220K No
Split on Refund:
8 No Single 85K Yes Entropy(Refund=Yes) = 0
9 No Married 75K No Entropy(Refund=No)
10 ? Single 90K Yes = -(2/6)log(2/6) – (4/6)log(4/6) = 0.9183
10

Entropy(Children)
Missing = 0.3 (0) + 0.6 (0.9183) = 0.551
value
Gain = 0.9  (0.8813 – 0.551) = 0.3303
© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 64
 Distribute Instances

Tid Refund Marital Taxable

Status Income Class
Tid Refund Marital Taxable
1 Yes Single 125K No Status Income Class
2 No Married 100K No
10 ? Single 90K Yes
3 No Single 70K No 10

4 Yes Married 120K No

Refund
5 No Divorced 95K Yes Yes No
6 No Married 60K No
Class=Yes 0 + 3/ 9 Class=Yes 2 + 6/ 9
7 Yes Divorced 220K No
Class=No 3 Class=No 4
8 No Single 85K Yes
9 No Married 75K No
10

Probability that Refund=Yes is 3/9

Refund
Yes No Probability that Refund=No is 6/9
Assign record to the left child with
Class=Yes 0 Cheat=Yes 2 weight = 3/9 and to the right child
Class=No 3 Cheat=No 4 with weight = 6/9

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 65

 Classify Instances

New record: Married Single Divorced Total

Tid Refund Marital Taxable
Status Income Class Class=No 3 1 0 4
11 No ? 85K ?
10

Class=Yes 6/9 1 1 2.67

Total 3.67 2 1 6.67

Refund
Yes
No
NO MarSt
Single,
Married Probability that Marital Status
Divorced
= Married is 3.67/6.67
TaxInc NO
< 80K > 80K Probability that Marital Status
={Single,Divorced} is 3/6.67
NO YES

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 66

 Other Issues

 Data Fragmentation
 Search Strategy
 Expressiveness
 Tree Replication

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 67

 Data Fragmentation

 Number of instances gets smaller as you traverse

down the tree

 Number of instances at the leaf nodes could be

too small to make any statistically significant
decision

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 68

 Search Strategy

 Finding an optimal decision tree is NP-hard

 The algorithm presented so far uses a greedy,

top-down, recursive partitioning strategy to
induce a reasonable solution

 Other strategies?
– Bottom-up
– Bi-directional

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 69

 Expressiveness

 Decision tree provides expressive representation for

learning discrete-valued function
– But they do not generalize well to certain types of
Boolean functions
 Example: parity function:
– Class = 1 if there is an even number of Boolean attributes with truth
value = True
– Class = 0 if there is an odd number of Boolean attributes with truth
value = True
 For accurate modeling, must have a complete tree

 Not expressive enough for modeling continuous variables

– Particularly when test condition involves only a single
attribute at-a-time

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 70

 Decision Boundary
1

0.9

0.8
x < 0.43?

0.7
Yes No
0.6

y < 0.33?
y

0.5 y < 0.47?

0.4

0.3
Yes No Yes No

0.2
:4 :0 :0 :4
0.1 :0 :4 :3 :0
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

x
• Border line between two neighboring regions of different classes is
known as decision boundary
• Decision boundary is parallel to axes because test condition involves
a single attribute at-a-time

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 71

 Oblique Decision Trees

x+y<1

Class = + Class =

• Test condition may involve multiple attributes

• More expressive representation
• Finding optimal test condition is computationally expensive

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 72

 Tree Replication
P

Q R

S 0 Q 1

0 1 S 0

0 1

• Same subtree appears in multiple branches

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 73

 Model Evaluation

 Metrics for Performance Evaluation

– How to evaluate the performance of a model?

 Methods for Performance Evaluation

– How to obtain reliable estimates?

 Methods for Model Comparison

– How to compare the relative performance
among competing models?

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 74

 Model Evaluation

 Metrics for Performance Evaluation

– How to evaluate the performance of a model?

 Methods for Performance Evaluation

– How to obtain reliable estimates?

 Methods for Model Comparison

– How to compare the relative performance
among competing models?

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 75

 Metrics for Performance
Evaluation

 Focus on the predictive capability of a model

– Rather than how fast it takes to classify or
build models, scalability, etc.
 Confusion Matrix:

PREDICTED CLASS
Class=Yes Class=No
a: TP (true positive)
b: FN (false negative)
Class=Yes a b
ACTUAL c: FP (false positive)

CLASS Class=No c d
d: TN (true negative)

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 76

 Metrics for Performance
Evaluation…

PREDICTED CLASS
Class=Yes Class=No

Class=Yes a b
ACTUAL (TP) (FN)
CLASS Class=No c d
(FP) (TN)

 Most widely-used metric:

ad TP  TN
Accuracy  
a  b  c  d TP  TN  FP  FN
© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 77
 Limitation of Accuracy

 Consider a 2-class problem

– Number of Class 0 examples = 9990
– Number of Class 1 examples = 10

 If model predicts everything to be class 0,

accuracy is 9990/10000 = 99.9 %
– Accuracy is misleading because model does
not detect any class 1 example

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 78

 Cost Matrix

PREDICTED CLASS

C(i|j) Class=Yes Class=No

Class=Yes C(Yes|Yes) C(No|Yes)

ACTUAL
CLASS Class=No C(Yes|No) C(No|No)

C(i|j): Cost of misclassifying class j example as class i

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 79

 Computing Cost of Classification

Cost PREDICTED CLASS

Matrix
C(i|j) + -
ACTUAL
+ -1 100
CLASS
- 1 0

Model M1 PREDICTED CLASS Model M2 PREDICTED CLASS

+ - + -
ACTUAL ACTUAL
+ 150 40 + 250 45
CLASS CLASS
- 60 250 - 5 200

Accuracy = 80% Accuracy = 90%

Cost = 3910 Cost = 4255
© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 80
 Cost vs Accuracy

Count PREDICTED CLASS Accuracy is proportional to cost if

1. C(Yes|No)=C(No|Yes) = q
Class=Yes Class=No
2. C(Yes|Yes)=C(No|No) = p
Class=Yes a b
ACTUAL N=a+b+c+d
CLASS Class=No c d

Accuracy = (a + d)/N

Cost PREDICTED CLASS

Cost = p (a + d) + q (b + c)
Class=Yes Class=No
= p (a + d) + q (N – a – d)
Class=Yes p q = q N – (q – p)(a + d)
ACTUAL
CLASS Class=No = N [q – (q-p)  Accuracy]
q p

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 81

 Cost-Sensitive Measures
a
Precision (p) 
a c
a
Recall (r) 
a b
2rp 2a
F - measure (F)  
r  p 2a  b  c
 Precision is biased towards C(Yes|Yes) & C(Yes|No)
 Recall is biased towards C(Yes|Yes) & C(No|Yes)
 F-measure is biased towards all except C(No|No)
wa w d
Weighted Accuracy  1 4

wa wbwcw d 1 2 3 4

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 82

 Model Evaluation

 Metrics for Performance Evaluation

– How to evaluate the performance of a model?

 Methods for Performance Evaluation

– How to obtain reliable estimates?

 Methods for Model Comparison

– How to compare the relative performance
among competing models?

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 83

 Methods for Performance
Evaluation

 How to obtain a reliable estimate of

performance?

 Performance of a model may depend on other

factors besides the learning algorithm:
– Class distribution
– Cost of misclassification
– Size of training and test sets

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 84

 Learning Curve

 Learning curve shows

how accuracy changes
with varying sample size
 Requires a sampling
schedule for creating
learning curve:
 Arithmetic sampling
(Langley, et al)
 Geometric sampling
(Provost et al)

Effect of small sample size:

- Bias in the estimate
- Variance of estimate

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 85

 Methods of Estimation
 Holdout
– Reserve 2/3 for training and 1/3 for testing
 Random subsampling
– Repeated holdout
 Cross validation
– Partition data into k disjoint subsets
– k-fold: train on k-1 partitions, test on the remaining one
– Leave-one-out: k=n
 Stratified sampling
– oversampling vs undersampling
 Bootstrap
– Sampling with replacement

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 86

 Model Evaluation

 Metrics for Performance Evaluation

– How to evaluate the performance of a model?

 Methods for Performance Evaluation

– How to obtain reliable estimates?

 Methods for Model Comparison

– How to compare the relative performance
among competing models?

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 87

 ROC (Receiver Operating
Characteristic)

 Developed in 1950s for signal detection theory to

analyze noisy signals
– Characterize the trade-off between positive
hits and false alarms
 ROC curve plots TP (on the y-axis) against FP
(on the x-axis)
 Performance of each classifier represented as a
point on the ROC curve
– changing the threshold of algorithm, sample
distribution or cost matrix changes the location
of the point

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 88

 ROC Curve
- 1-dimensional data set containing 2 classes (positive and negative)
- any points located at x > t is classified as positive

At threshold t:
TP=0.5, FN=0.5, FP=0.12, FN=0.88
© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 89
 ROC Curve

(TP,FP):
 (0,0): declare everything
to be negative class
 (1,1): declare everything
to be positive class
 (1,0): ideal

 Diagonal line:
– Random guessing
– Below diagonal line:
 prediction is opposite of
the true class

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 90

 Using ROC for Model Comparison

 No model consistently
outperform the other
 M is better for
1
small FPR
 M is better for
2
large FPR
 Area Under the ROC
curve
 Ideal:
 Area = 1
 Random guess:
 Area = 0.5

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 91

 How to Construct an ROC curve

• Use classifier that produces

Instance P(+|A) True Class
posterior probability for each
1 0.95 +
test instance P(+|A)
2 0.93 +
3 0.87 - • Sort the instances according
4 0.85 - to P(+|A) in decreasing order
5 0.85 -
• Apply threshold at each
6 0.85 +
unique value of P(+|A)
7 0.76 -
8 0.53 + • Count the number of TP, FP,
9 0.43 -
10 0.25 + TN, FN at each threshold
• TP rate, TPR = TP/(TP+FN)
• FP rate, FPR = FP/(FP + TN)
© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 92
 How to construct an ROC curve
Class + - + - - - + - + +
P
Threshold 0.25 0.43 0.53 0.76 0.85 0.85 0.85 0.87 0.93 0.95 1.00
>= TP 5 4 4 3 3 3 3 2 2 1 0

FP 5 5 4 4 3 2 1 1 0 0 0

TN 0 0 1 1 2 3 4 4 5 5 5

FN 0 1 1 2 2 2 2 3 3 4 5

TPR 1 0.8 0.8 0.6 0.6 0.6 0.6 0.4 0.4 0.2 0

FPR 1 1 0.8 0.8 0.6 0.4 0.2 0.2 0 0 0

ROC Curve:

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 93

 Test of Significance

 Given two models:

– Model M1: accuracy = 85%, tested on 30 instances
– Model M2: accuracy = 75%, tested on 5000 instances

 Can we say M1 is better than M2?

– How much confidence can we place on accuracy of
M1 and M2?
– Can the difference in performance measure be
explained as a result of random fluctuations in the test
set?

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 94

 Confidence Interval for Accuracy

 Prediction can be regarded as a Bernoulli trial

– A Bernoulli trial has 2 possible outcomes
– Possible outcomes for prediction: correct or wrong
– Collection of Bernoulli trials has a Binomial distribution:
 x  Bin(N, p) x: number of correct predictions
 e.g: Toss a fair coin 50 times, how many heads would turn up?
Expected number of heads = Np = 50  0.5 =
25

 Given x (# of correct predictions) or equivalently,

acc=x/N, and N (# of test instances),

Can we predict p (true accuracy of model)?

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 95
 Confidence Interval for Accuracy

 Area = 1 - 
For large test sets (N > 30),
– acc has a normal distribution
with mean p and variance
p(1-p)/N

acc  p
P( Z  Z )
p (1  p ) / N
 /2 1  / 2

1   Z/2 Z1-  /2

 Confidence Interval for p:

2 N acc  Z  Z  4 N acc  4 N acc
2 2 2

p  /2  /2

2( N  Z ) 2

 /2

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 96

 Confidence Interval for Accuracy

 Consider a model that produces an accuracy of

80% when evaluated on 100 test instances:
– N=100, acc = 0.8 1- Z
– Let 1- = 0.95 (95% confidence)
0.99 2.58
– From probability table, Z/2=1.96
0.98 2.33
N 50 100 500 1000 5000 0.95 1.96

p(lower) 0.670 0.711 0.763 0.774 0.789 0.90 1.65

p(upper) 0.888 0.866 0.833 0.824 0.811

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 97

 Comparing Performance of 2
Models

 Given two models, say M1 and M2, which is

better?
– M1 is tested on D1 (size=n1), found error rate = e1
– M2 is tested on D2 (size=n2), found error rate = e2
– Assume D1 and D2 are independent
– If n1 and n2 are sufficiently large, then
e1 ~ N 1 ,  1 
e2 ~ N 2 ,  2 
e (1  e )
– Approximate: ˆ  n
i i
i

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 98

 Comparing Performance of 2
Models

 To test if performance difference is statistically

significant: d = e1 – e2
– d ~ N(dt,t) where dt is the true difference
– Since D1 and D2 are independent, their variance adds
up:

    ˆ  ˆ
2

t 1
2

2
2 2

1
2

e1(1  e1) e2(1  e2)

 
n1 n2

– At (1-) confidence level, d d Z ˆ

t  /2 t

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 99

 An Illustrative Example

 Given: M1: n1 = 30, e1 = 0.15

M2: n2 = 5000, e2 = 0.25
 d = |e2 – e1| = 0.1 (2-sided test)

0.15(1  0.15) 0.25(1  0.25)

ˆ   0.0043
30 5000
d

 At 95% confidence level, Z/2=1.96

d 0.100 1.96  0.0043 0.100 0.128

t

=> Interval contains 0 => difference may not be

statistically significant
© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 100
 Comparing Performance of 2
Algorithms

 Each learning algorithm may produce k models:

– L1 may produce M11 , M12, …, M1k
– L2 may produce M21 , M22, …, M2k
 If models are generated on the same test sets
D1,D2, …, Dk (e.g., via cross-validation)
– For each set: compute dj = e1j – e2j
– dj has mean dt and variance t
k
– Estimate:  (d j  d) 2

ˆ  2 j 1

k (k  1)
t

d d t ˆ
t 1  , k  1 t

© Tan,Steinbach, Kumar Introduction to Data Mining 4/18/2004 101