Database Reference
In-Depth Information
Considering our tourists scenario and its set of daily trajectories of persons
moving in Paris with a GPS, the following behavior definition can be used to
separate the trajectories of tourists from the trajectories of other persons:
Tourist behavior : A daily trajectory shows the Tourist behavior if: Its
Begin point P1 is a place of kind “Accommodation,” it makes at least
one stop in a in a place of kind “Museum” or “TouristAttraction,”
it makes one stop in a in a place of kind “EatingPlace,” and its End
point is in the same P1 place as its Begin point.
This Tourist behavior is a semantic behavior. An example of spatio-temporal
behavior, always for the tourists scenario, is the LongTrajectory behavior defined
as: the duration of the trajectory is greater than 14 hours. The number of behaviors
that can be defined is unlimited. For example, “Going from the Place de la
Concorde to the Champs Elysees” and “Going from the Place de la Concorde
to Place de la Madeleine” could be semantic behaviors of interest for travel
agencies organizing tourist tours in Paris. Trajectories showing these behaviors
would also qualify as showing the more generic semantic behavior “Going from
a tourist spot to a commercial area.”
Interesting behaviors can be inferred using various methods for extracting
useful knowledge from trajectory data sets: data mining methods are discussed
in Chapters 6 and 7 of this topic, and visual analytics methods are discussed
in Chapter 8 . The most common outputs of these methods are clustering (tra-
jectories are grouped into classes that share some common characteristics) and
behaviors/patterns (describing the characteristics shared by significant groups
of trajectories).
Alternatively, an application manager interested in application-specific trends
canmanually define behaviors a priori. Back to our tourists scenario, a large num-
ber of behaviors can be manually predefined, each one targeting the identification
of a subset of the population moving in Paris: Tourist behavior, OfficeWorker
behavior, Housewife behavior, and so on. In some smaller-scale applications the
number of interesting behaviors may be small enough to be exhaustively defined.
Moving object database query languages, like the one presented in Chapters 3
and 12 of this topic, can be very effective for searching the trajectories that
comply with a given behavior.
Researchers have defined generic families of behaviors that rely on the con-
stancy/variation of some given characteristic of the trajectory (e.g., same direc-
tion or speed for a while) or on the similarity or correlation of the values of
some characteristic of a group of trajectories (e.g., proximity for a while). For
example, potentially interesting features in the shape or combinations of shapes
of trajectory traces have lead to the definition of a number of spatio-temporal
Search WWH ::




Custom Search