Information Technology Reference
In-Depth Information
are related and organized within a specific abstract control structure. Each rule
describes the concept in a recursive way by using:
-
A composition hierarchy: this is specified by the set of sub-concepts directly
composing the concept and their own composition hierarchies.
-
A set of constraints and conditions to be satisfied by the composing sub-
concepts, and all the relationships among them and with the sub-concepts
of the hierarchy.
A formalism for the specification of the recognition rules is given by
Attributed
Grammars
[13] for their expressiveness regarding the specification of the hierar-
chy, the constraints and relationships, as is well-known for the specification of
programming languages.
A production rule of the grammar specifies: a set
sub-concept
of terminal
and non-terminal symbols which represent the set of sub-concepts forming the
concept represented by the lhs symbol
concept
.
The set
condition
represent the relationships and constraints that must be
fulfilled by the sub-concepts forming the concept, in order to be recognized as a
valid instance.
The set
AttributionRule
of the production assigns values to the attributes
of the recognized concept utilizing the values of attributes of the composing
sub-concepts.
The syntax of a production rule is as follows:
Rule =
rule
concept
→
composition
{
subconcept
}
condition
[
local
LocalAttributes ]
{
Condition
}
attribution
{
AttributionRule
}
LocalAttributes =
attribute
:Type
{
}
attribute
:Type
∈ N
subconcept
∈ N ∪ T
attribute
∈ A
Condition
∈ C
AttributionRule
concept
∈ R
The Recognition Process.
The PDG information, together with syntactical in-
formation can be produced as a set of Prolog facts representing the
Abstract
Program Representation
. The hierarchical parsing process that do the recogni-
tion, is performed by an
Inferential Engine
that applies the production rules of
Search WWH ::
Custom Search