Database Reference
In-Depth Information
improve the accuracy of integration by making these semantic differences explicit in
ontologies, so the impact on a user's application can be seen. Furthermore, the links
and overlaps between the two datasets can be specified exactly, in a more detailed
fashion than just making the two categories of Field equivalent, so one query can be
carried out on the combination of the two datasets and achieve more accurate results.
2.4.2 D
ata
r
epurposIng
A second benefit of semantic technologies is in the reuse of data for purposes other
than the reason for which it was originally collected. For example, a mapping com-
pany might collect information about the spatial extents of objects such as sluice
gates and weirs to produce cartographic maps and provide spatial mapping data.
An organization reusing that data will have its own tasks to carry out and may need
to interpret the terminology differently. For example, the organization tasked with
flood risk management would have a need to identify all the possible flood defenses
in an area. A simple ontology can link the two views on the world by specifying
how the concepts used in one can be related to the concepts used by the other. In our
simple example, the statements: “A Sluice Gate is a kind of Flood Defense” and
“A Weir is a kind of Flood Defense” enable the topographic objects Sluice gate and
Weir to be connected to the world of flood defense. This in turn enables the Flood
Defense agency to query the mapping data for flood defenses, even though “flood
defense” was not a term explicit in map ontology data, and the mapping company
could not have predicted in advance during its data capture process all the possible
categories that its customers might in the future want to identify in its data.
2.4.3 D
ata
C
olleCtIon
, C
lassIfICatIon
,
anD
Q
ualIty
C
ontrol
Semantic technologies benefit the internal business processes of organizations that
collect or publish their own content. An ontology provides a very explicit data spec-
ification, which assists surveyors, other GI data collectors, or automatic collection
mechanisms in their task of identifying the information that needs to be added to or
modified in the database and assists with quality control. For example, a surveyor
who comes across a small watercourse might need to specify whether it was a bourn
or a stream. Many cases might be borderline, and an ontology can help with ensur-
ing data quality. For those designing the data specification, an ontological approach
provides a logical framework for addressing questions like, “What categories of
object do I need to collect information on?” “How do those categories relate to each
other?” “What instructions should I give to the Data Collection department on how
to identify and differentiate between Category A and Category B?”
The process of constructing an ontology to describe a data specification also
helps the organization clarify to itself exactly why it is capturing the data and
whether it is really needed. For example, a common mistake in traditional classifi-
cation systems is to confuse or conflate different axes of categorization. It is very
common for land use classification (the purpose or function of the land) to be mixed
up with a description of what the object is, so that a land use classification might
be Football Stadium (what something is) rather than its use or purpose: playing
