In Korean, there are word ambiguities because a word is constituted by one or more morpheme(s). Previous researches of parsing in Korean have ignored word ambiguities or used part-of-speech tagger. This propose a parsing model which does not ign... In Korean, there are word ambiguities because a word is constituted by one or more morpheme(s). Previous researches of parsing in Korean have ignored word ambiguities or used part-of-speech tagger. This propose a parsing model which does not ignore word ambiguities, using dependency grammar. To reduce ambiguities, syntactic and semantic techniques like chunking, sub-categorization, and argument information of verbs are used. Especially, the dictionary of argument information of verbs is utilized by Korean Word (KorLex). The experimental result shows that the parser based on our model outperforms other parsers in terms of recall rate. The recall of our parser is 94.42%. The precision of our parser is 79.63%, slightly lower than other parsers'. To improve the precision, probabilistic information, or N-Best algorithm will be used for the next studies. ，韩语论文题目，韩语论文范文 |
