Database Reference
In-Depth Information
the set of behaviors and the problem is to express these behaviors in terms of
movement characteristics to be used for searching a trajectory database. Here,
database approaches such as the one presented in Chapter 3 are suitable. A
more challenging issue arises when no behaviors are known a priori. How can
we learn potentially meaningful behaviors from trajectory analyses? Techniques
for this kind of research typically include data mining, machine learning, and
knowledge extraction in general, as well as visualization.
There are research efforts aiming at defining behaviors in a given domain in
a more abstract and generic way, for example, not for the purpose of a specific
application. These behaviors stem, for example, from an observation of possible
spatio-temporal configurations of moving objects and are assumed to be relevant
to a variety of applications. Other research aims at defining an ontology of all the
behaviors. We presented a set of basic concepts regarding behaviors. Chapters 6
and 7 develop a more detailed discussion on behaviors (called patterns).
1.7 Bibliographic Notes
Background knowledge on spatial, temporal, and spatio-temporal data descrip-
tion and management is largely covered by the literature. The well-known
ChoroChronos topic written by Koubarakis et al. ( 2003 ) reports the outcomes
from an early European project on spatio-temporal databases. Guting and
Schneider ( 2005 ) is an excellent reference topic on a formally sound approach
to moving object management. This approach, built on abstract data types, is
described in Chapter 3 of the present topic. Finally, a conceptual perspective
on spatio-temporal data modeling and manipulation is provided in Parent et al.
( 2006 ).
Most of the trajectory issues discussed in this chapter were first addressed in
Giannotti and Pedreschi ( 2008 ), a topic produced by the European GeoPKDD
project on privacy-preserving techniques for trajectory mining. In this topic, the
chapters on “Basic Concepts of Mobility Data” and “Trajectory Data Models”
nicely complement the content of our chapter.
The conceptual approach that has been very inspirational in writing this
chapter was published in a journal by Spaccapietra et al. ( 2008 ). This paper
develops a comprehensive view on trajectories from a conceptual data modeling
perspective. It introduces the concept of semantic trajectories and of segmented
trajectories, namely using Stop and Move episodes. Many further papers on
trajectory analysis stem from a similar approach.
Trajectory behaviors have been extensively addressed. Dodge et al. ( 2008 )
is one of few contributions that aim at proposing a taxonomy of behaviors for
raw trajectories. The authors studied the literature on data mining and visual
analysis dealing with movement data and they collected definitions of various
Search WWH ::




Custom Search