The advent of the high-speed computer with its enormous storage
capabilities enabled statisticians as well as researchers from the
different topics of life sciences to apply mul- tivariate statistical
procedures to large data sets to explore their structures. More and
more, methods of graphical representation and data analysis are used for
investigations. These methods belong to a topic of growing popUlarity,
known as "exploratory data analysis" or EDA. In many applications, there
is reason to believe that a set of objects can be clus- tered into
subgroups that differ in meaningful ways. Extensive data sets, for
example, are stored in clinical cancer registers. In large data sets
like these, nobody would ex- pect the objects to be homogeneous. The
most commonly used terms for the class of procedures that seek to
separate the component data into groups are "cluster analysis" or
"numerical taxonomy". The origins of cluster analysis can be found in
biology and anthropology at the beginning of the century. The first
systematic investigations in cluster analysis are those of K. Pearson in
1894. The search for classifications or ty- pologies of objects or
persons, however, is indigenous not only to biology but to a wide
variety of disciplines. Thus, in recent years, a growing interest in
classification and related areas has taken place. Today, we see
applications of cluster analysis not only to. biology but also to such
diverse areas as psychology, regional analysis, marketing research,
chemistry, archaeology and medicine.