Grid-clustering techniques are different from other clustering techniques in that they partition space instead of sets of points.

(a) How does this affect such techniques in terms of the description of the
resulting clusters and the types of clusters that can be found?
(b) What kind of cluster can be found with grid-based clusters that cannot
be found by other types of clustering approaches? (Hint: See Exercise
20 in Chapter 8, page 564.)

(a) In grid-based clustering, the clusters are described in terms of collec-
tions of adjacent cells. In some cases, as in CLIQUE, a more compact

description is generated. In any case, the description of the clusters is
given in terms of a region of space, not a set of objects. (However, such
a description can easily be generated.) Because it is necessary to work
in terms of rectangular regions, the shapes of non-rectangular clusters
can only be approximated. However, the groups of adjacent cells can
have holes.
(b) Typically, grid-based clustering techniques only pay attention to dense
regions. However, such techniques could also be used to identify sparse
or empty regions and thus find patterns of the absence of points. Note,
however, that this would not be appropriate for a sparse data space.