Lecture3 2020classification PDF

Classification: Basic Concepts and Decision
Trees
A programming task
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.
Illustrating Classification Task
Tid Attrib1 Attrib2 Attrib3 Class
1 Yes Large 125K No
2 No Medium 100K No
3 No Small 70K No
4 Yes Medium 120K No
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
10
Apply
Tid Attrib1 Attrib2 Attrib3 Class Model
11 No Small 55K ?
12 Yes Medium 80K ?
13 Yes Large 110K ?
14 No Small 95K ?
15 No Large 67K ?
10
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
Classification Using Distance
• Place items in class to which they are
“closest”.
• Must determine distance between an item
and a class.
• Classes represented by
– Centroid: Central value.
– Medoid: Representative point.
– Individual points
• Algorithm: KNN
Classification Techniques
• Decision Tree based Methods
• Rule-based Methods
• Memory based reasoning
• Neural Networks
• Naïve Bayes and Bayesian Belief Networks
• Support Vector Machines
A first example
Database of 20,000 images of handwritten digits, each
labeled by a human
[28 x 28 greyscale; pixel values 0-255; labels 0-9]
Use these to learn a classifier which will label digit-images automatically…

The learning problem
Input space X = {0,1,…,255}784
Output space Y = {0,1,…,9}
Training set Learning Classifier

(x1, y1), …, (xm, ym) Algorithm f: X ! Y
m = 20000
To measure how good f is: use a test set

[Our test set: 100 instances of each digit.]
A possible strategy
Input space X = {0,1,…,255}784
Output space Y = {0,1,…,9}
Treat each image as a point in 784-dimensional Euclidean space
To classify a new image: find its nearest neighbor in the database (training set) and
return that label
f = entire training set + search engine

K Nearest Neighbor (KNN):
• Training set includes classes.
• Examine K items near item to be classified.
• New item placed in class with the most
number of close items.
• O(q) for each tuple to be classified. (Here q
is the size of the training set.)
KNN
Nearest neighbor
Image to label Nearest neighbor
Overall:
error rate = 6%
(on test set)
Question: what is
the error rate
for random
guessing?
What does it get wrong?
Who knows… but here’s a hypothesis:
Each digit corresponds to some connected region of R784. Some of the regions come
close to each other; problems occur at these boundaries.
R1 R2
e.g. a random point in

this ball has only a
70% chance of
being in R2
Nearest neighbor: pros and cons
Pros
Simple
Flexible
Excellent performance on a wide range of tasks
Cons
Algorithmic: time consuming – with n training
points in Rd, time to label a new
point is O(nd)
Statistical: memorization, not learning!
no insight into the domain
would prefer a compact classifier
Prototype selection
A possible fix: instead of the entire training set, just keep a “representative
sample”
Voronoi
cells
“Decision boundary”
Prototype selection
A possible fix: instead of the entire training set, just keep a “representative
sample”
Voronoi
cells
“Decision boundary”
How to pick prototypes?
They needn’t be actual data points
Idea 2: one prototype per class: mean of training points
Examples:
Error = 23%
Postscript: learning models
Batch learning On-line learning
Training data See a new point x
Learning
Algorithm predict label
Classifier f See y
Update
test classifer
Example of a Decision Tree
Splitting Attributes
Tid Refund Marital Taxable
Status Income Cheat
1 Yes Single 125K No

2 No Married 100K No Refund
Yes No
3 No Single 70K No
4 Yes Married 120K No NO MarSt
5 No Divorced 95K Yes Single, Divorced Married
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

Another Example of Decision Tree
MarSt Single,
Married Divorced
Status Income Cheat
NO Refund
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
Decision Tree Classification Task
1 Yes Large 125K No
2 No Medium 100K No
3 No Small 70K No
5 No Large 95K Yes
6 No Medium 60K No
9 No Medium 75K No
10 No Small 90K Yes
10
Apply Decision
Model Tree
11 No Small 55K ?
12 Yes Medium 80K ?
13 Yes Large 110K ?
14 No Small 95K ?
15 No Large 67K ?
10
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
Test Data
Refund Marital Taxable
Status Income Cheat
No Married 80K ?
Refund 10
Yes No
NO MarSt
TaxInc NO
< 80K > 80K
NO YES
Test Data
Status Income Cheat
No Married 80K ?
Refund 10
Yes No
NO MarSt
TaxInc NO
< 80K > 80K
NO YES
Test Data
Status Income Cheat
No Married 80K ?
Refund 10
Yes No
NO MarSt
TaxInc NO
< 80K > 80K
NO YES
Test Data
Status Income Cheat
No Married 80K ?
Refund 10
Yes No
NO MarSt
TaxInc NO
< 80K > 80K
NO YES
Test Data
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
Decision Tree Classification Task
1 Yes Large 125K No
2 No Medium 100K No
3 No Small 70K No
5 No Large 95K Yes
6 No Medium 60K No
9 No Medium 75K No
10 No Small 90K Yes
10
Apply Decision
Model Tree
11 No Small 55K ?
12 Yes Medium 80K ?
13 Yes Large 110K ?
14 No Small 95K ?
15 No Large 67K ?
10
Decision Tree Induction
• Many Algorithms:
– Hunt’s Algorithm (one of the earliest)
– CART
– ID3, C4.5
– SLIQ,SPRINT
General Structure of Hunt’s Algorithm
• Let Dt be the set of training Status Income Cheat
records that reach a node t 1 Yes Single 125K No
• General Procedure: 2 No Married 100K No
– If Dt contains records that 3

4
No
Yes
Single
Married
70K
120K
No
belong the same class yt, then t

No
is a leaf node labeled as yt 6 No Married 60K No
– If Dt is an empty set, then t is a 7 Yes Divorced 220K No
leaf node labeled by the 8 No Single 85K Yes
default class, yd 9 No Married 75K No
– If Dt contains records that 10 No Single 90K Yes
belong to more than one class,

10
use an attribute test to split Dt

the data into smaller subsets.
Recursively apply the ?
procedure to each subset.
Hunt’s Algorithm
Status Income Cheat

Refund
Don’t 2 No Married 100K No
Yes No
Cheat 3 No Single 70K No
Don’t Don’t
4 Yes Married 120K No
Cheat Cheat
6 No Married 60K No
Refund Refund 8 No Single 85K Yes
Yes No Yes No 9 No Married 75K No
Don’t Marital 10 No Single 90K Yes
Don’t Marital Cheat
10
Cheat Status Status

Single, Single,
Married Married
Divorced Divorced
Don’t Taxable Don’t

Cheat Cheat
Cheat Income
< 80K >= 80K
Don’t Cheat
Cheat
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
Tree Induction
• Issues
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
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 OR CarType
{Sports, {Family,
Luxury} {Family} Luxury} {Sports}
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}
• What about this split?
Size
{Small,
Large} {Medium}
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
Splitting Based on Continuous Attributes
Tree Induction
• Issues
How to determine the Best Split
Before Splitting: 10 records of class 0,
10 records of class 1
Which test condition is the best?

How to determine the Best Split
• Greedy approach:
– Nodes with homogeneous class distribution are
preferred
• Need a measure of node impurity:
Non-homogeneous, Homogeneous,
High degree of impurity Low degree of impurity
Measures of Node Impurity
• Gini Index
• Entropy
• Misclassification error
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
Measure of Impurity: GINI
• Gini Index for a given node t :
GINI (t )  1   [ p ( j | t )]2
j
(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
Examples for computing GINI
GINI (t )  1   [ p ( j | t )]2
j
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
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.
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
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 C1 3 1 C1 2 2
C2 4 1 1 C2 2 4 C2 1 5
Gini 0.393 Gini 0.400 Gini 0.419
Continuous Attributes: Computing Gini
Index
• Use Binary Decisions based on one value Status Income Cheat

• Several Choices for the splitting value 2 No Married 100K No
– Number of possible splitting values 3

4
No
Yes
Single
Married
70K
120K
No
No
= Number of distinct values 5 No Divorced 95K Yes
• Each splitting value has a count matrix

6 No Married 60K No
associated with it 8 No Single 85K Yes

9 No Married 75K No
– Class counts in each of the partitions, A < v 10 No Single 90K Yes
and A  v
10
• Simple method to choose best v

– For each v, scan the database to gather
count matrix and compute its Gini index
– Computationally Inefficient! Repetition of
work.
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
Sorted Values 60 70 75 85 90 95 100 120 125 220

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
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
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
Splitting Based on INFO...
• Information Gain:
 n 
 Entropy ( p )  
k
GAIN 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.
Splitting Based on INFO...
• Gain Ratio:
GAIN n n
 SplitINFO    log
k
GainRATIO 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
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
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
Comparison among Splitting Criteria
For a 2-class problem:
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) = 0.342
= 1 – (4/7)2 – (3/7)2
= 0.489
Tree Induction
• Issues
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)

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
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
Practical Issues of Classification
• Underfitting and Overfitting
• Missing Values
• Costs of Classification
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
Underfitting and Overfitting
Overfitting
Underfitting: when model is too simple, both training and test errors are large
Overfitting due to Noise
Decision boundary is distorted by noise point

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
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

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
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
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.
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).
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
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
A? PRUNE!
A1 A4
A2 A3
Class = 8 Class = 3 Class = 4 Class = 5

Yes Yes Yes Yes
Class = 4 Class = 4 Class = 1 Class = 1
No No No No
Examples of Post-pruning
Case 1:
– Optimistic error?
Don’t prune for both cases
C0: 11 C0: 2
C1: 3 C1: 4
– Pessimistic error?
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
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
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
2 No Married 100K No Class Class
3 No Single 70K No = Yes = No
Refund=Yes 0 3
Refund=No 2 4
Refund=? 1 0
6 No Married 60K 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
Distribute Instances
Status Income Class
1 Yes Single 125K No Status Income Class
2 No Married 100K No
10 ? Single 90K Yes
3 No Single 70K No 10

Refund
5 No Divorced 95K Yes Yes No
6 No Married 60K No
Class=Yes 0 + 3/9 Class=Yes 2 + 6/9
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
Classify Instances
New record: Married Single Divorce Total
d
Status Income Class Class=No 3 1 0 4
11 No ? 85K ? Class=Yes 6/9 1 1 2.67

10
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
Scalable Decision Tree Induction Methods
• SLIQ (EDBT’96 — Mehta et al.)

– Builds an index for each attribute and only class list and the current
attribute list reside in memory
• SPRINT (VLDB’96 — J. Shafer et al.)
– Constructs an attribute list data structure
• PUBLIC (VLDB’98 — Rastogi & Shim)
– Integrates tree splitting and tree pruning: stop growing the tree earlier
• RainForest (VLDB’98 — Gehrke, Ramakrishnan & Ganti)
– Builds an AVC-list (attribute, value, class label)
• BOAT (PODS’99 — Gehrke, Ganti, Ramakrishnan & Loh)
– Uses bootstrapping to create several small samples
Postscript: learning models
Batch learning On-line learning
Training data See a new point x
Learning
Algorithm predict label
Classifier f See y
Update
test classifer
Preprocessing step
Generative and Discriminative
Classifiers
Generative vs. Discriminative Classifiers
Training classifiers involves estimating f: X  Y, or P(Y|X)
Discriminative classifiers (also called ‘informative’ by

Rubinstein&Hastie):
1. Assume some functional form for P(Y|X)
2. Estimate parameters of P(Y|X) directly from training data
Generative classifiers
1. Assume some functional form for P(X|Y), P(X)
2. Estimate parameters of P(X|Y), P(X) directly from training data
3. Use Bayes rule to calculate P(Y|X= xi)
Bayes Formula
Generative Model
• Color
• Size
• Texture
• Weight
• …
Discriminative Model
• Logistic Regression
• Color
• Size
• Texture
• Weight
• …
Comparison
• Generative models
– Assume some functional form for P(X|Y), P(Y)
– Estimate parameters of P(X|Y), P(Y) directly from
training data
– Use Bayes rule to calculate P(Y|X= x)
• Discriminative models
– Directly assume some functional form for P(Y|X)
– Estimate parameters of P(Y|X) directly from
training data
Probability Basics
• Prior, conditional and joint probability for random
variables
– Prior probability: P(X)
– Conditional probability: P(X1 |X2 ), P(X2 |X1 )
– Joint probability: X  (X1 , X2 ), P(X)  P(X1 ,X2 )
– Relationship: P(X1 ,X2 )  P(X2 |X1 )P(X1 )  P(X1 |X2 )P(X2 )
– Independence: P(X2 |X1 )  P(X2 ), P(X1 |X2 )  P(X1 ), P(X1 ,X2 )  P(X1 )P(X2 )
• Bayesian Rule
P(X|C)P(C) Likelihood Prior
P(C|X)  Posterior
P(X) Evidence
111
Probability Basics
• Quiz: We have two six-sided dice. When they are tolled, it could end up
with the following occurance: (A) dice 1 lands on side “3”, (B) dice 2 lands
on side “1”, and (C) Two dice sum to eight. Answer the following
questions:
1) P ( A )  ?
2) P(B)  ?
3) P(C)  ?
4) P ( A | B )  ?
5) P ( C | A )  ?
6) P ( A , B )  ?
7) P ( A , C )  ?
8) Is P ( A , C ) equals P(A)  P(C) ?
112
Probabilistic Classification
• Establishing a probabilistic model for classification
– Discriminative model
P(C|X) C  c1 ,  , cL , X  (X1 ,  , Xn )
P(c1 |x) P(c2 |x) P(cL |x)


Discriminative
Probabilistic Classifier

x1 x2 xn
x  ( x1 , x2 ,  , xn )
113
• Establishing a probabilistic model for classification
(cont.)
– Generative model
P(X|C) C  c1 ,  , cL , X  (X1 ,  , Xn )
P(x|c1 ) P(x|c2 ) P(x|cL )
Generative Generative Generative

Probabilistic Model Probabilistic Model  Probabilistic Model
for Class 1 for Class 2 for Class L
  
x1 x2 xn x1 x2 xn x1 x2 xn
x  ( x1 , x2 ,  , xn )
114
• MAP classification rule
– MAP: Maximum A Posterior
– Assign x to c* if
P(C  c* |X  x)  P(C  c|X  x) c  c* , c  c1 ,  , cL
• Generative classification with the MAP rule
– Apply Bayesian rule to convert them into posterior probabilities
P(X  x|C  ci )P(C  ci )
P(C  ci |X  x) 
P(X  x)
 P(X  x|C  ci )P(C  ci )
for i  1,2,  , L
– Then apply the MAP rule
115
Naïve Bayes
• Bayes classification
P(C|X)  P(X|C)P(C)  P(X1 ,  , Xn |C)P(C)
Difficulty: learning the joint probability P(X1 ,  , Xn |C)
• Naïve Bayes classification
– Assumption that all input attributes are conditionally
independent!
P(X1 , X2 ,  , Xn |C)  P(X1 |X2 ,  , Xn ; C)P(X2 ,  , Xn |C)
 P(X1 |C)P(X2 ,  , Xn |C)
 P(X1 |C)P(X2 |C)    P(Xn |C)
– MAP classification rule: for x  ( x1 , x2 ,  , xn )

[P( x1 |c* )    P( xn |c* )]P(c* )  [P( x1 |c)    P( xn |c)]P(c), c  c* , c  c1 ,  , cL
116
Naïve Bayes
• Naïve Bayes Algorithm (for discrete input attributes)
– Learning Phase: Given a training set S,
For each target valueof ci (ci  c1 ,  , cL )
Pˆ (C  ci )  estimateP(C  ci ) with examplesin S;
For everyattributevalue xjk of each attributeXj ( j  1,  , n; k  1,  , N j )
Pˆ (Xj  xjk |C  ci )  estimateP(Xj  xjk |C  ci ) with examplesin S;
Output: conditional probability tables; for Xj , Nj Lelements
– Test Phase: Given an unknown instance X  ( a ,  , a ) ,
1 n
Look up tables to assign the label c* to X’ if
[Pˆ (a1 |c* )    Pˆ (an |c* )]Pˆ (c* )  [Pˆ (a1 |c)    Pˆ (an |c)]Pˆ (c), c  c* , c  c1 ,  , cL
117
Example
• Example: Play Tennis
118
Example
• Learning Phase
Outlook Play=Yes Play=No Temperature Play=Yes Play=No
Sunny 2/9 3/5
Hot 2/9 2/5
Overcast 4/9 0/5
Mild 4/9 2/5
Rain 3/9 2/5
Cool 3/9 1/5
Humidity Play=Yes Play=No
Wind Play=Yes Play=No
High 3/9 4/5

Strong 3/9 3/5
Normal 6/9 1/5
Weak 6/9 2/5
P(Play=Yes) = 9/14 P(Play=No) = 5/14
119
Example
• Test Phase
– Given a new instance,
x’=(Outlook=Sunny, Temperature=Cool, Humidity=High, Wind=Strong)
– Look up tables
P(Outlook=Sunny|Play=Yes) = 2/9 P(Outlook=Sunny|Play=No) = 3/5
P(Temperature=Cool|Play=Yes) = 3/9 P(Temperature=Cool|Play==No) = 1/5
P(Huminity=High|Play=Yes) = 3/9 P(Huminity=High|Play=No) = 4/5
P(Wind=Strong|Play=Yes) = 3/9 P(Wind=Strong|Play=No) = 3/5
– MAP rule
P(Yes|x’): [P(Sunny|Yes)P(Cool|Yes)P(High|Yes)P(Strong|Yes)]P(Play=Yes) =
0.0053
P(No|x’): [P(Sunny|No) P(Cool|No)P(High|No)P(Strong|No)]P(Play=No) =
0.0206
Given the fact P(Yes|x’) < P(No|x’), we label x’ to be “No”. 120

Example
• Test Phase
– Given a new instance,
x’=(Outlook=Sunny, Temperature=Cool, Humidity=High, Wind=Strong)
– Look up tables
P(Outlook=Sunny|Play=Yes) = 2/9 P(Outlook=Sunny|Play=No) = 3/5
P(Temperature=Cool|Play=Yes) = 3/9 P(Temperature=Cool|Play==No) = 1/5
P(Huminity=High|Play=Yes) = 3/9 P(Huminity=High|Play=No) = 4/5
P(Wind=Strong|Play=Yes) = 3/9 P(Wind=Strong|Play=No) = 3/5
– MAP rule
P(Yes|x’): [P(Sunny|Yes)P(Cool|Yes)P(High|Yes)P(Strong|Yes)]P(Play=Yes) =
0.0053
P(No|x’): [P(Sunny|No) P(Cool|No)P(High|No)P(Strong|No)]P(Play=No) =
0.0206
Given the fact P(Yes|x’) < P(No|x’), we label x’ to be “No”. 121

Relevant Issues
• Violation of Independence Assumption
– For many real world tasks, P(X1,  ,Xn|C)  P(X1|C)    P(Xn|C)
– Nevertheless, naïve Bayes works surprisingly well anyway!
• Zero conditional probability Problem
– If no example contains the attribute value Xj  ajk , Pˆ(Xj  ajk |C  ci )  0
– In this circumstance, Pˆ(x1|ci )  Pˆ(ajk|ci ) Pˆ(xn|ci )  0during test
– For a remedy, conditional probabilities estimated with
n  mp
Pˆ ( X j  a jk |C  ci )  c
nm
nc : number of training examples for which X j  a jk and C  ci
n : number of training examples for which C  ci
p : prior estimate (usually, p  1 / t for t possible values of X j )
m : weight to prior (number of " virtual" examples, m  1)
122
Relevant Issues
• Continuous-valued Input Attributes
– Numberless values for an attribute
– Conditional probability modeled with the normal distribution
1  ( X j   ji )2 
Pˆ ( X j |C  ci )  exp  
2  ji  2 ji 
2

 ji : mean (avearage) of attribute values X j of examples for which C  ci
 ji : standard deviation of attribute values X j of examples for which C  ci
– Learning Phase: for X  ( X1 ,  , Xn ), C  c1 ,  , cL

Output: n L normal distributions and P(C  ci ) i  1,  , L
– Test Phase: for X  (X1 ,  , Xn )
• Calculate conditional probabilities with all the normal distributions
• Apply the MAP rule to make a decision
123
Conclusions
• Naïve Bayes based on the independence assumption
– Training is very easy and fast; just requiring considering each
attribute in each class separately
– Test is straightforward; just looking up tables or calculating
conditional probabilities with normal distributions
• A popular generative model
– Performance competitive to most of state-of-the-art classifiers even
in presence of violating independence assumption
– Many successful applications, e.g., spam mail filtering
– A good candidate of a base learner in ensemble learning
– Apart from classification, naïve Bayes can do more…
124

Lecture3 2020classification PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lecture3 2020classification PDF

Uploaded by

Copyright:

Available Formats

Classification: Basic Concepts and Decision

• Classifying credit card transactions

• Classifying secondary structures of protein

• Categorizing news stories as finance,

[28 x 28 greyscale; pixel values 0-255; labels 0-9]

Use these to learn a classifier which will label digit-images automatically…

Training set Learning Classifier

To measure how good f is: use a test set

Treat each image as a point in 784-dimensional Euclidean space

f = entire training set + search engine

e.g. a random point in

Idea 2: one prototype per class: mean of training points

Training data See a new point x

1 Yes Single 125K No

Training Data Model: Decision Tree

records that reach a node t 1 Yes Single 125K No

• General Procedure: 2 No Married 100K No

– If Dt contains records that 3

belong the same class yt, then t

– If Dt contains records that 10 No Single 90K Yes

belong to more than one class,

use an attribute test to split Dt

1 Yes Single 125K No

Cheat Status Status

Don’t Taxable Don’t

• Depends on number of ways to split

• Binary split: Divides values into two subsets.

– Binary Decision: (A < v) or (A  v)

Which test condition is the best?

Node N1 Node N2 Node N3 Node N4

C0 N10 C0 N20 C0 N30 C0 N40

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

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

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

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

where, ni = number of records at child i,

CarType CarType CarType

1 Yes Single 125K No

– Number of possible splitting values 3

• Each splitting value has a count matrix

associated with it 8 No Single 85K Yes

• Simple method to choose best v

Cheat No No No Yes Yes Yes No No No No

Sorted Values 60 70 75 85 90 95 100 120 125 220

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

– Measures homogeneity of a node.

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

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

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

Parent Node, p is split into k partitions;

Parent Node, p is split into k partitions

• Measures misclassification error made by a node.

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

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

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

• Stop expanding a node when all the records

• Early termination (to be discussed later)

• You can download the software from:

500 circular and 500

Decision boundary is distorted by noise point

• Training error no longer provides a good

• Need new ways for estimating errors

• For complex models, there is a greater chance