METHOD FOR AUTOMATICALLY EXTRACTING KOREAN FOREIGN WORD

摘要

PPROBLEM TO BE SOLVED: To automatically detect(extract) foreign words in a Korean corpus without making it necessary to execute the morphemic analysis processing of a translation corpus by using phonologically similar characteristics of both Japanese/Korean foreign words, and automatically extracting the foreign words in the Korean corpus with Japanese KATAKANA words as a clue. PSOLUTION: A Korean corpus are Roman character-converted, and a Roman character-converted text is divided into phrases according to a Korean word division writing rule, and postpositional particles and affixes connected to the tail of words are removed from the divided phrases so that the words can be obtained, and the words whose similarity with Roman character-converted Japanese KATAKANA words are extracted as Korean foreign words from the words being foreign word candidates by removing any phrase including alphabetic characters which are not used for foreign word notation and the words registered in an existing Korean dictionary. PCOPYRIGHT: (C)2005,JPO&NCIPI