CLUSTER ANALYSIS: ALGORITHMS AND ANALYSIS USING SAS

CLUSTER ANALYSIS:
ALGORITHMS AND ANALYSIS USING SAS

BY: AHMED ALDAHHAN
SUPERVISED BY: LECTURER JING XU
BIRKBECK UNIVERSITY OF LONDON
2013/2014

ABSTRACT
The scope of this paper is to provide an introduction to cluster analysis; by giving a general background for cluster analysis; and explaining the concept of cluster analysis and how the clustering algorithms work. A basic idea and the use of each clustering method will be described with its graphical features. Different clustering techniques are also explained with examples to get a better idea. The two main clustering techniques (Hierarchical and K-means Partitioning) are illustrated using a sample data set ‘IRIS FLOWER DATA
SET’ (1936), where a comparison of the two methods is made based on data suitability and model performance. TABLE OF CONTENTS
CHAPTER 1
1.0

Introduction …………………………………………………………………………………………………….. 5

1.1

Understanding Cluster Analysis ……………..……………………….……………………………….. 7

CHAPTER 2
2.0

Definitions …………………………………………………………………………..………………………..… 9

2.1

The Data Matrix ………………………………………………………..…….…………………………….… 9

2.2

The Proximity matrix ………………………………………………………………….……………………. 9

2.3

Similarity and Dissimilarity Matrices ………..………………..………………………………..…. 11

2.4

Different Types of Clusters ……………………………………………………………………..………. 11
2.4.1
2.4.2

Centre-Based cluster definition ……………………………………………………………… 12

2.4.3

Contiguity-Based Cluster Definition ………………………………………………….……. 13

2.4.4

Density-Based Clusters definition …………………………………………………..……… 13

2.4.5
2.5

Well-Separated cluster definition ………………………………………………………….. 11

Shared-Property ( Conceptual Clusters ) ………………………………………………... 14

Distance Matrix ………………………………………………………………………………………………. 14

2.6 Hierarchical Clustering …………………………………………….………………………………….….… 16
2.6.1

Agglomerative Hierarchical Clustering …………………………………………………... 16

2.6.2

Divisive Hierarchical Clustering

Cited: DEC 13