摘要 |
PURPOSE:To prevent the over-conversion of KANA (Japanese syllabary) into KANJI (chinese characters) by retrieving the KANA character strings emerging frequently out of a KANA character string store means and excluding them out of the KANJI conversion objects as a pre-stage where the KANA/KANJI conversion is performed after the retrieval of a word dictionary. CONSTITUTION:A character string written only in KANA characters and to be segmented is set and then the arrangements of KANA characters are sorted by the KANJI sounds against said character string. In case a KANJI sound plus KANJI sound, a KANJI sound plus KANA and KANA plus sounds excepting KANJI sounds are decided, the pre-processing is carried out with a HIRAGANA (cursive form of Japanese syllabary) character string (>=5 characters). That is, the corresponding KANA character string is extracted with the reference to a KANA character string memory 8. Then the existence of an additional work is checked and a character string to be retrieved by a dictionary is defined by the arrangement of the head and following KANJI sounds and KANA characters in case no additional word is decided. Thus the access frequency is decreased to the dictionary. In other words, both the processing load and processing time are reduced just by excluding a HIRAGANA character string (>=5 characters) out of the object of KANJI conversion.
|