국어 어휘 분석 말뭉치 구축과 활용

Abstract

The purpose of this paper is to introduce 'Korean word-tagged corpus' created by NAKL(the National Academy of the Korean Language). It can be one of the fundamental data of listing basic Korean vocabulary for beginners. Corpus-tagging is the process of attaching linguistic information to raw corpus. Ambiguity of a word can be found in any level; morphological, syntactic, semantic, pragmatic level and so on. Corpus-tagging which can reduce the number of possible interpretations helps to resolve semantic level of the ambiguity in natural language. Word-tagged corpus is a kind of the semantic annotated corpus. Source of word tags is 'Standard Korean Dictionary'. In a word entries in 'Standard Korean Dictionary' corresponds to word tags of the corpus. There are two stages in tagging word information. At first grammatical annotation can be performed automatically by a morphological analyzer. And it is needed to correct the errors of tagger and attach the tag for word annotation manually. We used semi-automatic system 'Wordanal' for word tagging. It links corpus DB to dictionary DB. To analyze the word-tagged corpus, we got various kinds of list and frequency about characters, syllables, words, phrase, etc. It was shown at appendixes. It can be used for framing language policy, teaching language, compiling dictionary, processing language data, etc.