Information Technology Reference
In-Depth Information
Chapter 3
Movement Mining
With ever increasing volumes and complexity of spatio-temporal information,
knowledge discovery in databases and its best known step data mining, have rapidly
gained importancewithinGeography andGIScience. Analyzing spatio-temporal data
first of all means structuring data, then extracting relevant spatial patterns and rules
and providing decision makers with enriched information and condensed knowl-
edge rather than flooding them with raw data. Movement patterns, for example,
represent such sought-for high-level process knowledge derived from low-level tra-
jectory data. This second chapter introducing the research field of Computational
Movement Analysis (CMA) reviews research on several aspects of mining move-
ment data, including the conceptualization and formalization of movement patterns
and the development of algorithms for their detection, the computing of trajectory
similarity, and methods for visualization-based exploratory analysis of movement
data.
1
Overarching research objectives
. The research summarized in this chapter con-
tributes to the following overarching research objectives of computational movement
analysis.
•
Illustrate to what respect the conceptual underpinning and toolset of data mining
suits the specific requirements of computational movement analysis.
•
Exemplify how geographically-informed movement mining contributes to a better
understanding of movement processes.
•
Promote a more thorough attitude towards evaluation of proposed movement min-
ing methods.
1
Whereas this chapter discusses movement mining in conventional omniscient and centralized
information systems or databases, the following Chap.
4
discusses the rather peculiar case where
data mining is performed in decentralized systems such as geosensor networks. Even though most
of the work summarized in Chap.
4
nominally also proposes data mining techniques, its theoretical
underpinning in decentralized spatial computing justifies a separate chapter focusing on decentral-
ized movement analysis alone.