Databases Reference
In-Depth Information
generate a quality, restricted set of candidate attributes against which the algorithm
for finding contextual candidate attributes can be tested.
3.4
Discussion
Contextual correspondences improve on noncontextual correspondences in two
main aspects. First, it allows the specification of alternative correspondences for
the same attribute, under different conditions. Second, its refined definition of a cor-
respondence allows it to connect attributes via correspondences in cases where a
common correspondence is too “weak” to be considered valid.
The way contextual correspondences are defined, they are deterministic, not
allowing a probabilistic interpretation of a correspondence. Therefore, contextual
correspondences are meant to resolve an issue of uncertainty by finding a more
refined, yet deterministic, correspondence. It is worth noting, however, that an intro-
duction of stochastic analysis already exists in the form of statistical significance,
which can be extended to handle probabilistic mappings as well.
4
Semantic Attribute Correspondences
Traditionally, attribute correspondences are limited to the ontological construct
known as
synonym
. The model of semantic matching, introduced in the realm of
the semantic Web [
Euzenat and Shvaiko 2007
], extends attribute correspondences
in that pairwise matching results in an ontological relationship (e.g., equivalence
and subsumption) between attributes. Therefore, a semantic attribute correspon-
dence is a triplet of the form .a
i
;a
j
;r/,wherea
i
and a
j
are attributes and r is
an ontological relationship. The S-Match system [
Giunchiglia et al. 2005
]defines
four types of such relationships, equivalence (
D
), subsumption (
v
), disjointness
(
)in
Magnani
et al. [
2005
]. We shall use S-Match to illustrate the semantic extension to attribute
correspondences.
S-Match separates labels from concepts. A label (referred to as
concept at label
in
Magnani et al.
[
2005
]) represents its meaning in the real world while a concept is
a composite term that considers the label within its path in the semantic graph (e.g.,
ontology, classification tree). To illustrate the difference, consider the case study
and Fig.
3.3
. The relational model schema can be interpreted as a semantic graph, in
which a relation name is linked with its attribute through a part-of relationship and a
primary key is related to its foreign key through a special, foreign-key, relationship.
We note here that in
Giunchiglia et al.
[
2005
] the type of the semantic link is ignored
when computing semantic attribute correspondences. In our case study example,
the label
city
appears multiple times and is considered to have the same real world
meaning(s). However, the concept
city
is different when part of
HotelInfo
or part of
Subway
.
?
), and unknown (idk). The latter was referred to as intersection (
\
