Database Reference
In-Depth Information
time
Y
X
Figure 6.2 Visual example of a moving cluster over three time units.
found, formed by three objects, and any pair of consecutive clusters shares two
of the three objects. This way, moving clusters that last a long time might even
start from a set of objects and end in a completely different (possibly disjoint)
set. In our example, only one object permanently belongs to the moving cluster.
In some sense, the pattern is not strictly related to a population that generates
it. The purpose of the pattern becomes to describe phenomena that happen in
the population, not to find a group of individuals that do something peculiar
consistently together.
One element of rigidity that affects both the patterns illustrated so far is the
fact that they describe continuous portions of time. For instance, if a herd that
usually moves compactly gets dispersed for a short time (for instance, due to an
attack by predators) and later becomes compact again, both flocks and moving
clusters will generally result into two different and disconnected patterns - the
before
and the
after
the temporary dispersion. One possible way to avoid this loss
of information consists of allowing
gaps
in the patterns, that is, a pattern involves
a set of timestamps that are not necessarily consecutive. In the literature we can
find a solution of this kind, known as
swarm patterns
. Swarms are a general form
of patterns that generalize flocks and moving clusters, as any spatial clustering
method can be applied at the level of a single timestamp, and then spatial clusters
belonging to different timestamps are linked (in case they share an appropriate
fraction of population) regardless of their temporal distance.
6.2.2 Using Relative Time
In some contexts, the moving objects we are examining might act in a similar
way, even if they are not spatially located together. For instance, similar daily
routines might lead several individuals to drive their cars along the same routes,
even if they leave home at very different hours of the day. Or, tourists who visit
a city on different days of the year might actually visit it in the same way -
for instance, by visiting the same places in the same order and spending there
