Concept Parsing Algorithms (CPA)

Concept Parsing Algorithms (CPA)

DOI: 10.4018/978-1-5225-2176-1.ch003


This chapter describe Concept Parsing Algorithms (CPA), a novel methodology of using text analysis tools for discovery of ‘building blocks' of concepts, with semantic searches of the full text of potentially relevant documents in relevant knowledge domains, for lexical labels of concepts in controlled vocabularies. The meaning of lexical label of a super-ordinate concept C' in a sublanguage with controlled vocabulary is encoded in a set that contains three sets of building blocks: Ci (set of co-occurring sub-ordinate concepts); Rj (set of relations); and Lk (set of linguistic elements/descriptors).
Chapter Preview

Main Focus Of The Chapter

Issues, Controversies, Problems

In the following discussion we define as ‘super-ordinate’ a concept at the focus of interest, and denote it C’. Equation (3) is a set-theoretic definition of the super-ordinate concept C’ in terms of its three building blocks, each containing a specific type of features/descriptors; this equation can be used as a generic format of Concept Parsing Algorithms (CPA) that guide the unpacking of C’ into its component parts:

C’ = {[Ci], [Rj], [Lk]} (3)

The meaning of a lexical label of a super-ordinate concept C’ in a sublanguage with controlled vocabulary is encoded in a set that contains three building blocks (Shafrir & Etkind, 2005; 2006); these are the sets:

  • [Ci] = set of co-occurring sub-ordinate concepts [C1, C2, …, Cm]

  • [Rj] = set of relations [R1, R2, …, Rn]

  • [Lk] = set of linguistic elements (descriptors) [L1, L2, …, Lp]

Sets [Ci], [Rj] and [Lk] in Equation (3) can be characterized by the following (non-exhaustive) list of descriptors:

Set [Ci] of Co-occurring Concepts:

  • 1.

    The set must contain at least two co-occurring concepts (i >= 2; cannot be an empty set)

  • 2.

    Each concept has a unique lexical label which acts as a proper name; no synonyms are allowed

  • 3.

    Each concept occurs unconditionally

  • 4.

    Co-occurring concepts are unranked

  • 5.

    No metric is available for comparing co-occurring concepts

Set [Rj] of Relations:

  • The set may be empty (j =0)

  • A relation does not have a unique lexical label and may accept synonyms

  • A relation may be between co-occurring concepts, or between co-occurring concepts and the super-ordinal concept

  • A relation between two concepts is unconditional

  • Relations are unranked

  • No metric is available for comparing relations

Set [Lk] of Linguistic Elements:

  • The set may not be empty (k >=1)

  • Linguistic elements must obey syntactic, morphological, and grammatical rules of the language

Complete Chapter List

Search this Book: