A Korean Noun Semantic Hierarchy (Wordnet) Construction
18123Authors
nan
Abstract
Since thesaurus is used as a knowledge resource in many natural language processing systems, it is very useful and necessary for the high quality systems, especially for dealing with semantics. In this paper, we introduce a semi-automatic method for the construction of Korean noun semantic hierarchy by utilizing a monolingual MRD and an existing thesaurus. 1�
Keywords
Full Content
A Korean Noun Semantic Hierarchy (Wordnet) Construction Juho Lee, Koaunghi Un, Hee-Sook Bae and.Key-Sun Choi KORTERM, Dept. of EECS, KAIST 373-1 Guseong-dong, Yuseong-gu Taejon 305-701, Korea {mywork, koaunghi, elle, kschoi}@world.kaist.ac.kr Abstract Since thesaurus is used as a knowledge resource in many natural language processing systems, it is very useful and necessary for the high quality systems, especially for dealing with semantics. In this paper, we introduce a semi-automatic method for the construction of Korean noun semantic hierarchy by utilizing a monolingual MRD and an existing thesaurus. 1� Introduction Thesaurus 1 or wordnet takes too much time and effort to construct them manually. In this paper, we introduce a semi-automatic method for the construction of Korean noun semantic hierarchy with lexical mapping to each noun's sense. In this method, an MRD (Hangeul Society, ed. 1997) and an existing translated thesaurus (Ikehara, et al. 1997) are used. By assigning the semantic category of the existing thesaurus to each sense2 of the nouns in MRD, we combine these two resources andproduce an expanded l exicalized n oun s emantic h ierarchy. T he s emantic c ategory i s assigned first and manual correction is performed in post-processing, so semantic hierarchy is constructed with relatively high accuracy and small effort. 2 Related Works There were researches on construction of a WordNet using existing WordNet and MRD. In EuroWordNet project, a multilingual database for several European languages was built based on English WordNet. For example, Spanish WordNet was constructed using English WordNet (Atserias, et a1. 1997, Farreres, et a1. 1998). They used a Spanish monolingual dictionary and bilingual dictionaries for the construction. They supposed three methods and combined those methods in order to get high coverage. For Korean, Moon (1996) used hypernym information of a Korean dictionary and combined it with Korean translation of the English WordNet. The manual pruning was done during the construction. There was another research for the construction of Korean WordNet based on the English WordNet (Lee et al. 2000). They used a bilingual dictionary to link the senses of Korean nouns to the synsets of English WordNet. They applied six heuristics to word sense disambiguation and combined each heuristic with decision tree. This work is different from preceding works in three points; 1) our Korean noun semantic hierarchy is a lexical map including hierarchy and other relationships, not a WordNet; 2) since it is based on a Japanese thesaurus and Korean MRD, we have an advantage of similarity. between two languages; 3) we link lexical units, elements of thesaurus, with each sense of nouns in MRD. 1 "Thesauri are based on concepts and they show relationships among terms. Relationships commonly expressed in a thesaurus include hierarchy, equivalence (synonymy), and association or relatedness." [.. "WordNet structures concepts and terms not as hierarchies but as a network or a web" (Hodge 2000) 2 Because the terms, "sense" and "signification" are differently treated by linguists, we clarify that the "sense" indicates the diverse realizations of "signification" corresponding to a lexical unit in …
Content truncated to 500 words
Labeled Text
Since thesaurus is used as a knowledge resource in many natural language processing systems, it is very useful and necessary for the high quality systems, especially for dealing with semantics. In this paper, we introduce a semi-automatic method for the construction of Korean noun semantic hierarchy by utilizing a monolingual MRD and an existing thesaurus. 1�
Publication Details
- Paper ID:
- 18123
- Year:
- 2001
- Publication:
- TIPSTER
- Publication Type:
- conf
- DOI:
- nan