Automatic Construction of a Concept Hierarchy from Coordinated Noun Phrases

다국어 초록 (Multilingual Abstract)

Noun phrase coordination is an extremely productive phenomenon. Based on an observation that conjuncts tend to denote semantically related concepts, we collect four hundred thousand pairs of conjuncts from the British National Corpus, in an attempt to build an is-a hierarchy of English noun concepts. The modifiedness patterns of the two words in these pairs point to three distinct semantic relations: sibling, cousin, and ancestor-or-ancestor`s-sibling. The process of finding them and how these pairs are used to motivate groups of quasi-synonyms and then to locate the hypernyms are discussed.

참고문헌 (Reference)

1 "WordNet: An Electronic Lexical Database" MITPress 1998

2 "Using information content to evaluate semantic similarity" 448-453, 1995

3 "The descent of Hierar-chy, and selection in relational semantics" 247-254, 2002

4 "Ontologies on demand? {A description of the state-of-the-art, applications, challenges and trends for ontology-learning from text" 57 (57): 315-320, 2006

5 "Nouns in WordNet FivePapers on WordNet" 1993

6 "Linear time algorithms for testing the satis abilityof propositional Horn formulae Journal of Logic Programming" 1984

7 "Journal of Linguistics" 1985

8 "Generalizing case frames using a thesaurus and the MDL Principle" 24 (24): 217-244, 1998

9 "Finding Structure in Language Centre for CognitiveScience University of Edinburgh" Finch 1993

10 "English verbs as a semantic net" 1993

11 "Contexts of antonymous adjectives" 1989

12 "Automatic retrieval and clustering of similar words" 98 : 768-773, 1998

13 "Automatic construction of a hypernym-labeled noun hier-archy from text" 2000

14 "Automatic acquisition of hyponyms from large text corpora InProcedings of the Fourteenth International Conference on Computational Linguistics" 1992

15 "Accurate unlexicalized parsing" 2003