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 89

Figure 6. MIEL++ architecture

ontology. The query is translated by each subsys-
tem's wrapper into a query expressed in the query
language of the subsystem: an SQL query in the
relational subsystem (see Buche & al. (2005) for
more details about the SQL subsystem), a SPARQL
query in the XML/RDF subsystem (SPARQL is
a query language recommended by the W3C to
query RDF data sources http://www.w3.org/TR/
rdf-sparql-query/). Finally, the global answer to
the query is the union of the local results of the
two subsystems, which are ordered according to
their relevance to the query selection criteria.
In this section, we are interested in the XML/
RDF subsystem which allows the end-user to query
RDF annotations of Web data tables represented
in XML documents. A MIEL++ query is asked in
a view which corresponds to a given relation of
the ontology. A view is characterized by its set of
queryable attributes and by its actual definition.
Each queryable attribute corresponds to a type of
the relation represented by the view. The concept
of view must be understood with the meaning of
the relational database model: it permits to hide the
complexity of the querying in a given subsystem
to the end-user. The end-user uses a view to build

his query. In the XML/RDF subsystem, a view
is defined by means of a SPARQL generic query
where the SELECT clause contains the queryable
attributes and the WHERE clause contains the
definition of the view.

Example 4
Let us consider the view V associated with the
relation ContaminationRange of domain Food-
Product, Contaminant, SamplesNumberTotal and
of range ContaminationLevel. The SPARQL query
associated with V is presented in Figure 7, the
where part of the query being shown graphically
for readability reasons.
A MIEL++ query is an instance of a given view
specified by the end-user, by choosing, among
the set of queryable attributes of the view, which
are the conjunctive selection attributes and their
corresponding searched values, which are the
projection attributes and which is the minimal
threshold on the pertinence score associated with
the relation represented by the view. In a MIEL++
query, the end-user can express preferences in