148
Omar Boucelma, Mehdi Essid, and Yassine Lassoued
rules allow the mediator (i) to select the appropriate sources and data for the user’s
query and (ii) to rewrite the user’s query into queries over the data sources’ schemas.
Mapping rules are expressed in a language that supports restrictions and conversion
functions.
In order to facilitate query rewriting and to ensure system extensibility (addition
and deletion of data sources), a mapping rule describes a part (subtree) of the global
schema according to only one data source schema. Combining information of various
data sources is performed automatically by means of a global key mechanism. The
idea behind this mechanism is to relate each attribute of a class to a key on which
it depends. Finally, we use quality aggregation operations in order to estimate the
quality of integrated data sets.
7.6.4 Quality Mediator
In order to define the quality mediator, we extended Lenzerini’s definition for data
integration systems [24], in adding a quality model, and a global key mechanism.
Definition 7.1 (Quality Mediation System). A quality mediation system is defined
as a tuple I = (G, S, D, Q, key), where G is a global schema, S = {S i , i ∈ [1, . . . , n]}
is a set of source schemas, D = {M i , i ∈ [1, . . . , n]} is the set of their respective
descriptions, that is for each i ∈ [1, . . . , n], M i is the set of mapping rules between S i
and G. The term Q denotes a quality model and key denotes a function that associates
to each class of the global schema a global key.
Definition 7.2 (Global Key). Given a class C of a global schema G, a global key is
an attribute or a set of attributes of class C allowing identification of class instances
(i.e. on which the values of the remaining attributes depend), whatever the (local)
data source they are extracted from.
The key mechanism we are using here is a simplified version of XML keys defined
in [7] for XML documents and is similar to the one described in [1]. A global key
plays the role of an object identifier. We suppose that each class C has a global key.
As an example, the key attribute of class Road is its geometric property. Note that
the id attribute cannot be used as a global key because it allows one to identify only
objects of the same data source; thus it is called a local key [1].
7.6.5 Quality Model
A quality model allows the representation of various data quality parameters. It is
mainly composed of quality criteria called quality parameters or quality elements.
Several geographical quality models have been proposed by standard or public institutions such as FGDC [13], the ISO/TC 211 model [22], or the IGN [11] that we
adopted as our quality reference model.
Omar Boucelma, Mehdi Essid, and Yassine Lassoued
rules allow the mediator (i) to select the appropriate sources and data for the user’s
query and (ii) to rewrite the user’s query into queries over the data sources’ schemas.
Mapping rules are expressed in a language that supports restrictions and conversion
functions.
In order to facilitate query rewriting and to ensure system extensibility (addition
and deletion of data sources), a mapping rule describes a part (subtree) of the global
schema according to only one data source schema. Combining information of various
data sources is performed automatically by means of a global key mechanism. The
idea behind this mechanism is to relate each attribute of a class to a key on which
it depends. Finally, we use quality aggregation operations in order to estimate the
quality of integrated data sets.
7.6.4 Quality Mediator
In order to define the quality mediator, we extended Lenzerini’s definition for data
integration systems [24], in adding a quality model, and a global key mechanism.
Definition 7.1 (Quality Mediation System). A quality mediation system is defined
as a tuple I = (G, S, D, Q, key), where G is a global schema, S = {S i , i ∈ [1, . . . , n]}
is a set of source schemas, D = {M i , i ∈ [1, . . . , n]} is the set of their respective
descriptions, that is for each i ∈ [1, . . . , n], M i is the set of mapping rules between S i
and G. The term Q denotes a quality model and key denotes a function that associates
to each class of the global schema a global key.
Definition 7.2 (Global Key). Given a class C of a global schema G, a global key is
an attribute or a set of attributes of class C allowing identification of class instances
(i.e. on which the values of the remaining attributes depend), whatever the (local)
data source they are extracted from.
The key mechanism we are using here is a simplified version of XML keys defined
in [7] for XML documents and is similar to the one described in [1]. A global key
plays the role of an object identifier. We suppose that each class C has a global key.
As an example, the key attribute of class Road is its geometric property. Note that
the id attribute cannot be used as a global key because it allows one to identify only
objects of the same data source; thus it is called a local key [1].
7.6.5 Quality Model
A quality model allows the representation of various data quality parameters. It is
mainly composed of quality criteria called quality parameters or quality elements.
Several geographical quality models have been proposed by standard or public institutions such as FGDC [13], the ISO/TC 211 model [22], or the IGN [11] that we
adopted as our quality reference model.