
7.5 Fuzzy Clustering Concepts
■
223
clusters are distinctly arranged in the n-dimensional sample space. In
practice (that is, in reality), however, cluster perimeters are often quite
ambiguous. Data points in the sample space share attributes from one or
more clusters and should thus be placed in more than one cluster (albeit
with a degree of membership in one cluster that is higher or lower than
the degree of membership in another). This brings us to the focus of the
chapter: fuzzy clustering and classification.
7.5 Fuzzy Clustering Concepts
A cluster brings together instances in the data that share a common set
of attributes. For each of these clusters ...