Transcript Clustering

Segmentação (Clustering)
(baseado nos slides do
Han)
Non-supervised Learning:
Cluster Analysis

What is Cluster Analysis?

Types of Data in Cluster Analysis

A Categorization of Major Clustering Methods

Partitioning Methods

Hierarchical Methods

Summary
SAD Tagus 2004/05
H. Galhardas
What is Cluster
Analysis?

Cluster: a collection of data objects



Cluster analysis



Similar to one another within the same cluster
Dissimilar to the objects in other clusters
Grouping a set of data objects into clusters
Clustering is unsupervised classification: no
predefined classes
Typical applications


As a stand-alone tool to get insight into data distribution
As a preprocessing step for other algorithms
SAD Tagus 2004/05
H. Galhardas
General Applications of
Clustering


Pattern Recognition
Spatial Data Analysis





create thematic maps in GIS by clustering feature spaces
detect spatial clusters and explain them in spatial data
mining
Image Processing
Economic Science (especially market research)
WWW
Document classification
 Cluster Weblog data to discover groups of similar access
patterns
H. Galhardas
SAD Tagus 2004/05

Examples of Clustering
Applications

Marketing: Help marketers discover distinct groups in their
customer bases, and then use this knowledge to develop
targeted marketing programs

Land use: Identification of areas of similar land use in an
earth observation database

Insurance: Identifying groups of motor insurance policy
holders with a high average claim cost

City-planning: Identifying groups of houses according to
their house type, value, and geographical location

Earth-quake studies: Observed earth quake epicenters
should be clustered along continent faults
SAD Tagus 2004/05
H. Galhardas
What Is Good Clustering?

A good clustering method will produce high quality
clusters with

high intra-class similarity

low inter-class similarity

The quality of a clustering result depends on both the
similarity measure used by the method and its
implementation.

The quality of a clustering method is also measured by its
ability to discover some or all of the hidden patterns.
SAD Tagus 2004/05
H. Galhardas
Requirements of Clustering
in Data Mining

Scalability

Ability to deal with different types of attributes

Discovery of clusters with arbitrary shape

Minimal requirements for domain knowledge to
determine input parameters

Able to deal with noise and outliers

Insensitive to order of input records

High dimensionality

Incorporation of user-specified constraints

Interpretability and usability
SAD Tagus 2004/05
H. Galhardas
Data Structures



x
 11
 ...
x
 i1
 ...

 xn1
Data matrix
(two modes)
Dissimilarity matrix

(one mode)
SAD Tagus 2004/05
... x
1f
... ...
...
...
x
if
...
... x
nf
... x 
1p 
... ... 
... x 
ip 
... ... 

... x 
np 
 0
 d(2,1)
0

 d(3,1) d ( 3,2) 0

:
:
 :
d ( n,1) d ( n,2) ...
H. Galhardas






... 0
Measure the Quality of
Clustering





Dissimilarity/Similarity metric: Similarity is expressed in
terms of a distance function, which is typically a metric:d(i, j)
There is a separate “quality” function that measures the
“goodness” of a cluster.
The definitions of distance functions are usually very
different for interval-scaled, boolean, categorical, ordinal and
ratio variables.
Weights should be associated with different variables based
on applications and data semantics.
It is hard to define “similar enough” or “good enough”

the answer is typically highly subjective.
SAD Tagus 2004/05
H. Galhardas
Type of data in clustering
analysis

Interval-scaled variables:

Binary variables:

Nominal, ordinal, and ratio variables:

Variables of mixed types:
SAD Tagus 2004/05
H. Galhardas
Interval-valued variables

Standardize data

Calculate the mean absolute deviation:
sf  1
n (| x1 f  m f |  | x2 f  m f | ... | xnf  m f |)
mf  1
where
n (x1 f  x2 f  ...  xnf )
.


Calculate the standardized measurement (z-score)
xif  m f
zif 
sf
Using mean absolute deviation is more robust
than using standard deviation
SAD Tagus 2004/05
H. Galhardas
Similarity and Dissimilarity
Between Objects

Distances are normally used to measure the
similarity or dissimilarity between two data objects

Some popular ones include: Minkowski distance
d (i, j)  q (| x  x |q  | x  x |q ... | x  x |q )
i1
j1
i2
j2
ip
jp
where i = (xi1, xi2, …, xip) and j = (xj1, xj2, …, xjp) are
two p-dimensional data objects, and q is a positive
integer

If q = 1, d is Manhattan distance
d (i, j) | x  x |  | x  x | ... | x  x |
i1 j1 i2 j 2
ip j p
SAD Tagus 2004/05
H. Galhardas
Similarity and Dissimilarity
Between Objects (Cont.)

If q = 2, d is Euclidean distance:
d (i, j)  (| x  x |2  | x  x |2 ... | x  x |2 )
i1
j1
i2
j2
ip
jp
Properties:

d(i,j)  0

d(i,i) = 0

d(i,j) = d(j,i)

d(i,j)  d(i,k) + d(k,j)
Also, one can use weighted distance,
parametric Pearson product moment
correlation, or otherH. disimilarity
measures
Galhardas
SAD Tagus 2004/05

Partitioning Algorithms:
Basic Concept

Partitioning method: Construct a partition of a
database D of n objects into a set of k clusters

Given a k, find a partition of k clusters that
optimizes the chosen partitioning criterion

Global optimal: exhaustively enumerate all partitions

Heuristic methods: k-means and k-medoids algorithms

k-means (MacQueen’67): Each cluster is represented
by the center of the cluster
k-medoids or PAM (Partition around medoids)
(Kaufman & Rousseeuw’87): Each cluster is
represented by one ofH.the
objects in the cluster
Galhardas
SAD Tagus 2004/05

The K-Means Clustering
Method

Given k, the k-means algorithm is
implemented in four steps:
1.
Partition objects into k nonempty subsets
2.
Compute seed points as the centroids of the
clusters of the current partition (the centroid is
the center, i.e., mean point, of the cluster)
3.
Assign each object to the cluster with the nearest
seed point
Go back to Step 2, stop when no more new
assignment
H. Galhardas
SAD Tagus 2004/05
4.
The K-Means Clustering
Method

Example
10
10
9
9
8
8
7
7
6
6
5
5
4
4
3
2
1
0
0
1
2
3
4
5
6
7
8
K=2
Arbitrarily choose K
object as initial
cluster center
9
10
Assign
each
objects
to most
similar
center
10
9
8
7
6
5
Update
the
cluster
means
3
2
1
0
0
1
2
3
4
5
6
7
8
9
10
3
2
1
0
0
1
2
3
4
5
6
reassign
10
10
9
9
8
8
7
7
6
6
5
5
4
2
1
0
0
1
2
3
4
5
6
7
8
H. Galhardas
7
8
9
10
reassign
3
SAD Tagus 2004/05
4
9
10
Update
the
cluster
means
4
3
2
1
0
0
1
2
3
4
5
6
7
8
9
10
Comments on the K-Means
Method
Strength Relatively efficient: O(tkn), where n is # objects, k is #
clusters, and t is # iterations. Normally, k, t << n.

Comparing: PAM: O(k(n-k)2 ), CLARA: O(ks2 + k(n-k))
Comment Often terminates at a local optimum. The global
optimum may be found using techniques such as: deterministic
annealing and genetic algorithms
Weakness

Applicable only when mean is defined, then what about categorical data?

Need to specify k, the number of clusters, in advance

Unable to handle noisy data and outliers

Not suitable to discover clusters with non-convex shapes
SAD Tagus 2004/05
H. Galhardas
Variations of the K-Means
Method


A few variants of the k-means which differ in

Selection of the initial k means

Dissimilarity calculations

Strategies to calculate cluster means
Handling categorical data: k-modes (Huang’98)

Replacing means of clusters with modes

Using new dissimilarity measures to deal with categorical objects

Using a frequency-based method to update modes of clusters

A mixture of categorical and numerical data: k-prototype method
SAD Tagus 2004/05
H. Galhardas
What is the problem of kMeans Method?

The k-means algorithm is sensitive to outliers !

Since an object with an extremely large value may substantially
distort the distribution of the data.

K-Medoids: Instead of taking the mean value of the object
in a cluster as a reference point, medoids can be used,
which is the most centrally located object in a cluster.
SAD Tagus 2004/05
10
10
9
9
8
8
7
7
6
6
5
5
4
4
3
3
2
2
1
1
0
0
1
2
3
4
5
6
7
8
9
10
H. Galhardas
0
0
1
2
3
4
5
6
7
8
9
10
The K-Medoids Clustering
Method

Find representative objects, called medoids, in
clusters

PAM (Partitioning Around Medoids, 1987)

starts from an initial set of medoids and iteratively replaces
one of the medoids by one of the non-medoids if it
improves the total distance of the resulting clustering

PAM works effectively for small data sets, but does not
scale well for large data sets
SAD Tagus 2004/05
H. Galhardas
Typical k-medoids algorithm
(PAM)
Total Cost = 20
10
10
10
9
9
9
8
8
8
Arbitrary
choose k
object as
initial
medoids
7
6
5
4
3
2
7
6
5
4
3
2
1
1
0
0
0
1
2
3
4
5
6
7
8
9
0
10
1
2
3
4
5
6
7
8
9
10
Assign
each
remainin
g object
to
nearest
medoids
7
6
5
4
3
2
1
0
0
K=2
Until no
change
10
3
4
5
6
7
8
9
10
10
Compute
total cost of
swapping
9
9
Swapping O
and Oramdom
8
If quality is
improved.
5
5
4
4
3
3
2
2
1
1
7
6
0
8
7
6
0
0
SAD Tagus 2004/05
2
Randomly select a
nonmedoid object,Oramdom
Total Cost = 26
Do loop
1
1
2
3
4
5
6
H. Galhardas
7
8
9
10
0
1
2
3
4
5
6
7
8
9
10
PAM (Partitioning Around
Medoids) (1987)

Use real object to represent the cluster
1.
Select k representative objects arbitrarily
2.
For each pair of non-selected object h and selected
object i, calculate the total swapping cost TCih
3.
For each pair of i and h,
4.
•
If TCih < 0, i is replaced by h
•
Then assign each non-selected object to the most
similar representative object
Repeat steps 2-3 until there is no change
SAD Tagus 2004/05
H. Galhardas
Cost function for k-medoids
SAD Tagus 2004/05
H. Galhardas
PAM Clustering
10
10
9
9
8
j
t
8
t
7
7
6
5
j
6
h
4
5
4
i
3
h
i
3
2
2
1
1
0
0
0
1
2
3
4
5
6
7
8
9
10
0
1
2
3
4
5
6
7
8
9
10
Cjih = 0
Cjih = d(j, h) - d(j, i)
10
10
9
9
8
8
h
7
7
j
6
6
i
5
5
i
4
4
t
3
2
2
1
1
j
h
t
3
0
0
0
0
1
SAD Tagus 2004/05
2
3
4
5
6
7
8
9
Cjih = d(j, t) - d(j, i)
1
2
3
4
5
6
7
8
9
10
H. Galhardas
Cjih = d(j, h) - d(j, t)
10
What is the problem
with PAM?

Pam is more robust than k-means in the
presence of noise and outliers because a
medoid is less influenced by outliers or other
extreme values than a mean

Pam works efficiently for small data sets but
does not scale well for large data sets.

O(k(n-k)2 ) for each iteration, where n is # of
data,k is # of cluster
SAD Tagus 2004/05
H. Galhardas
Summary





Cluster analysis groups objects based on their similarity
and has wide applications
Measure of similarity can be computed for various types
of data
Clustering algorithms can be categorized into partitioning
methods, hierarchical methods, density-based methods,
grid-based methods, and model-based methods
Outlier detection and analysis are very useful for fraud
detection, etc. and can be performed by statistical,
distance-based or deviation-based approaches
There are still lots of research issues on cluster analysis,
such as constraint-based clustering
SAD Tagus 2004/05
H. Galhardas
Other Classification
Methods

k-nearest neighbor classifier
case-based reasoning

Genetic algorithm

Rough set approach

Fuzzy set approaches

SAD Tagus 2004/05
H. Galhardas
Instance-Based
Methods

Instance-based learning:


Store training examples and delay the processing
(“lazy evaluation”) until a new instance must be
classified
Typical approaches



k-nearest neighbor approach
 Instances represented as points in a Euclidean
space.
Locally weighted regression
 Constructs local approximation
Case-based reasoning
 Uses symbolic representations and knowledgebased inference
SAD Tagus 2004/05
H. Galhardas
The k-Nearest Neighbor
Algorithm



All instances correspond to points in the
n-dimensional space.
The nearest neighbor are defined in
terms of Euclidean distance.
The target function could be discrete- or
real- valued.
SAD Tagus 2004/05
H. Galhardas
The k-Nearest Neighbor
Algorithm


For discrete-valued, the k-NN returns the
most common value among the k
training examples nearest to xq.
Vonoroi diagram: the decision surface
induced by 1-NN for a typical set of
training examples.
.
_
_
_
+
_
_
SAD Tagus 2004/05
.
+
+
xq
_
.
+
H. Galhardas
.
.
.
Discussion (1)

The k-NN algorithm for continuous-valued target
functions


Calculate the mean values of the k nearest neighbors
Distance-weighted nearest neighbor algorithm


Weight the contribution of each of the k neighbors
according to their distance to the query point xq,giving
greater weight to closer neighbors
1
w
Similarly, for real-valued target functions
d ( x , x )2
SAD Tagus 2004/05
q i
H. Galhardas
Discussion (2)


Robust to noisy data by averaging k-nearest
neighbors
Curse of dimensionality: distance between
neighbors could be dominated by irrelevant
attributes.

To overcome it, axes stretch or elimination of the least
relevant attributes.
SAD Tagus 2004/05
H. Galhardas
Bibliografia

Data Mining: Concepts and Techniques, J.
Han & M. Kamber, Morgan Kaufmann,
2001 (Sect. 7.7.1 e Cap. 8)
SAD Tagus 2004/05
H. Galhardas