Ontology-Based Integration of Heterogeneous, Incomplete and Imprecise Data Dedicated to a Decision Support System for Food Safety - Data Warehousing Design and Advanced Engineering Applications - page 88

Figure 5. RDF annotations of the first row of the Web data table presented in Table 1

of domain the symbolic types FoodProduct ,
Contaminant and Samples Total Number and of
range the numeric type Contamination Level has
been identified.
Figure 5 presents a part of the RDF descriptions
corresponding to the recognition of the relation
ContaminationRange in the first row of Table 1.
The first row (having the URI uriRow1 in the XML
document) is an instance of the Contamination-
Range relation recognized with a pertinence score
of 0.75. This pertinence score is computed by the
semantic annotation process as the proportion of
recognized types of the relation. It expresses the
degree of “certainty” associated with the relation
recognition. A part of the domain of the relation
presented in the example (typed by the OWL class
AssociatedKey ) is an instance of the symbolic
type FoodProduct (food1) and is annotated by a
discrete fuzzy set ( DFS1 ) which has a semantic of
similarity. It indicates the list of closest values of
the ontology ( Rice and Rice Flour ) compared to
the value Basmati Rice . The range of the relation
(typed by the OWL class AssociatedResult ) is an
instance of the numeric type ContaminationLevel
and is annotated by a continuous fuzzy set ( CFS1 )
which has a semantic of imprecision. It indicates
the possible contamination limits ([1.65, 1.95])

represented as the support and the kernel of the
fuzzy set.
The output of the @WEB system is an XML/
RDF data warehouse composed of a set of XML
documents which represent data tables and contain
their RDF annotations.

The Flexible Querying System

In order to deal with the data heterogeneity in the
CONTA sources, we propose to the end-user a
unified querying system, called MIEL++, which
permits to query simultaneously the CONTA local
relational database and the CONTA XML/RDF
data warehouse in a transparent way. The MIEL++
system is a flexible querying system relying on
a given domain ontology, the CONTA ontology.
It allows the end-user to retrieve the nearest data
stored in both sources corresponding to his/her
selection criteria: the CONTA ontology -more
precisely the type hierarchies- is used in order to
assess which data can be considered as “near” to
the user's selection criteria.
Figure 6 gives an overview of the MIEL++
architecture. When a query is asked to the MIEL++
system, that query is asked through a single graphi-
cal user interface, which relies on the domain