provided in this chapter. The utility of this technique to solve several benchmark
test problems is shown. Chapter 3 mainly describes the different types of similarity
measures developed in the literature for handling binary, categorical, ordinal, and
quantitative variables. It also contains a discussion on different normalization techniques.
Chapter 4 gives a broad overview of the existing clustering techniques and
their relative advantages and disadvantages. It also provides a detailed discussion of
several recently developed single- and multiobjective metaheuristic clustering techniques.
Chapter 5 presents symmetry-based distances and a genetic algorithm-based
clustering technique that uses this symmetry-based distance for assignment of points
to different clusters and for fitness computation. In Chap. 6, some symmetry-based
cluster validity indices are described. Elaborate experimental results are also presented
in this chapter. Application of these symmetry-based cluster validity indices
to segment remote sensing satellite images is also presented. In Chap. 7, some automatic
clustering techniques based on symmetry are presented. These techniques are
able to determine the appropriate number of clusters and the appropriate partitioning
from data sets having symmetric clusters. The effectiveness of these clustering techniques
is shown for many artificial and real-life data sets, including MR brain image
segmentation. Chapter 8 deals with an extension of the concept of point symmetry to
line symmetry-based distances, and genetic algorithm-based clustering techniques
using these distances. Results are presented for some artificial and real-life data sets.
A technique using the line symmetry-based distance for face detection from images
is also discussed in detail. Some multiobjective clustering techniques based on symmetry
are described in detail in Chap. 9. Three different clustering techniques are
discussed here. The first one assumes the number of clusters a priori. The second
and third clustering techniques are able to detect the appropriate number of clusters
from data sets automatically. The third one, apart from using the concept of symmetry,
also uses the concept of connectivity. A method of measuring connectivity
between two points in a cluster is described for this purpose. A connectivity-based
cluster validity index is also discussed in this chapter. Extensive experimental results
illustrating the greater effectiveness of the three multiobjective clustering techniques
over the single-objective approaches are presented for several artificial and real-life
data sets.
This book contains an in-depth discussion on clustering and its various facets.
In particular, it concentrates on metaheuristic clustering using symmetry as a similarity
measure with extensive real-life applications in data mining, satellite remote
sensing, MR brain imaging, gene expression data, and face detection. It is, in this
sense, a complete book that will be equally useful to the layman and beginner as to
an advanced researcher in clustering, being valuable for several reasons.
It includes discussions on traditional as well as some recent symmetry-based similarity
measurements. Existing well-known clustering techniques along with metaheuristic
approaches are elaborately described. Moreover, some recent clustering
techniques based on symmetry are described in detail. Multiobjective clustering is
another emerging topic in unsupervised classification. A multiobjective data clustering
technique is described elaborately in this book along with extensive experimental
results. A chapter of this book is wholly devoted to discussing some existing and