7 A Quality-enabled Spatial Integration System
149
Quality Parameters
The IGN quality model we have adopted is composed of the following parameters:
Q 0 = B b : Spatial cover
Q 1 = P 1 : Lineage
Q 2 = P 2 : Currentness
Q 3 = P 3 : Geometric accuracy
Q 4 = P 4 : Semantic accuracy
Q 5 = C 1 : Completeness
Q 6 = C 2 : Consistency
Each of the parameters Q i , i ∈ [0..6], is evaluated through a tuple of subparameters (Q
1
i , · · · , Q
k i
i ). For the sake of simplicity, we only consider the IGN subparameters listed in Table 7.1. Because Q 5 and Q 6 have a different behavior, they are
denoted respectively by C 1 and C 2 . In fact, as explained later, these are percentages,
which can be improved or degraded by combining data sets.
The quality model is attached to the global schema, each parameter applies to
one or more types in the global schema. Column 2 of Table 7.1 indicates to which
type of elements each parameter (or sub-parameter) is applicable. Each quality subparameter, when applied to a global schema element, is evaluated using a value of a
given type. The last column of Table 7.1 represents the type of each sub-parameter.
Quality Assertions
Definition 7.3 (Quality Assertion). A quality assertion κ associates to each quality
sub-parameter Q
j
i , i ∈ [0..6], j ∈ [1..k i ], and an element E of the global schema
(such as Q
j
i is applicable to E) a quality value κ(Q
j
i , E). For ordered quality subparameters, κ returns an interval (a rectangle in the case of the spatial cover),
whereas for others it returns a finite set of values.
Table 7.1. Quality parameters and sub-parameters
parameter applicable to
sub-parameter significance
type
Q 0
geographical classes Q
1
0 = B b
bounding box
boundingbox
Q 1 = P 1
geometric properties Q
1
1 = P
1
1
scale
float
classes, attributes
Q
2
1 = P
2
1
provider
string
Q
3
1 = P
3
1
source
string
Q
4
1 = P
4
1
source organization
string
Q 2 = P 2
classes, attributes
Q
1
2 = P
1
2
validation date
date
Q
2
2 = P
2
2
date of validity
date
Q
3
2 = P
3
2
expiry date
date
Q
4
2 = P
4
2
frequency of updates integer
Q 3 = P 3
geometric properties Q
1
3 = P
1
3
position accuracy
float
Q
2
3 = P
2
3
shape accuracy
float
Q 4 = P 4
attributes
Q
1
4 = P
1
4
semantic accuracy
enumeration/float
Q 5 = C 1
classes, attributes
Q
1
5 = C
1
1
deficit rate
percentage
Q
2
5 = C
2
1
surplus rate
percentage
Q 6 = C 2
classes
Q
1
6 = C
1
2
accordance rate
percentage
attributes
Q
2
6 = C
2
2
violation rate
percentage
149
Quality Parameters
The IGN quality model we have adopted is composed of the following parameters:
Q 0 = B b : Spatial cover
Q 1 = P 1 : Lineage
Q 2 = P 2 : Currentness
Q 3 = P 3 : Geometric accuracy
Q 4 = P 4 : Semantic accuracy
Q 5 = C 1 : Completeness
Q 6 = C 2 : Consistency
Each of the parameters Q i , i ∈ [0..6], is evaluated through a tuple of subparameters (Q
1
i , · · · , Q
k i
i ). For the sake of simplicity, we only consider the IGN subparameters listed in Table 7.1. Because Q 5 and Q 6 have a different behavior, they are
denoted respectively by C 1 and C 2 . In fact, as explained later, these are percentages,
which can be improved or degraded by combining data sets.
The quality model is attached to the global schema, each parameter applies to
one or more types in the global schema. Column 2 of Table 7.1 indicates to which
type of elements each parameter (or sub-parameter) is applicable. Each quality subparameter, when applied to a global schema element, is evaluated using a value of a
given type. The last column of Table 7.1 represents the type of each sub-parameter.
Quality Assertions
Definition 7.3 (Quality Assertion). A quality assertion κ associates to each quality
sub-parameter Q
j
i , i ∈ [0..6], j ∈ [1..k i ], and an element E of the global schema
(such as Q
j
i is applicable to E) a quality value κ(Q
j
i , E). For ordered quality subparameters, κ returns an interval (a rectangle in the case of the spatial cover),
whereas for others it returns a finite set of values.
Table 7.1. Quality parameters and sub-parameters
parameter applicable to
sub-parameter significance
type
Q 0
geographical classes Q
1
0 = B b
bounding box
boundingbox
Q 1 = P 1
geometric properties Q
1
1 = P
1
1
scale
float
classes, attributes
Q
2
1 = P
2
1
provider
string
Q
3
1 = P
3
1
source
string
Q
4
1 = P
4
1
source organization
string
Q 2 = P 2
classes, attributes
Q
1
2 = P
1
2
validation date
date
Q
2
2 = P
2
2
date of validity
date
Q
3
2 = P
3
2
expiry date
date
Q
4
2 = P
4
2
frequency of updates integer
Q 3 = P 3
geometric properties Q
1
3 = P
1
3
position accuracy
float
Q
2
3 = P
2
3
shape accuracy
float
Q 4 = P 4
attributes
Q
1
4 = P
1
4
semantic accuracy
enumeration/float
Q 5 = C 1
classes, attributes
Q
1
5 = C
1
1
deficit rate
percentage
Q
2
5 = C
2
1
surplus rate
percentage
Q 6 = C 2
classes
Q
1
6 = C
1
2
accordance rate
percentage
attributes
Q
2
6 = C
2
2
violation rate
percentage