| 摘要 |
PURPOSE:To reduce the number of words in a specific expression recognized as unknown words to improve the analysis precision by retrieving a dictionary with an expression as the retrieval key in the case of morpheme analysis which divides a character string consisting of at least two or more kinds of expression into word units and recognizing unknown words of the character string in the specific expression even at the time of failure of retrieval. CONSTITUTION:A KATAKANA(square form of Japanese syllabary) word recognizing part 8 converts the KATAKANA string detected by a KATAKANA string detecting part 7 to a HIRAGANA(cursive form of Japanese syllabary) string by a HIRAGANA conversion part 9. This HIRAGANA string is regarded as reading information of the KATAKANA string and is used as the retrieval key to retrieve word dictionary parts 2 and 11 by a reading retrieval part 10. When a word is found by retrieval, a KATAKANA word recognizing part 8 recognizes the word in a text string. |
