| 摘要 |
PURPOSE:To ensure a natural read-out system together with improvement of intelligibility by identifying a character string with an unknown word of three characters included in its immediately preceding word in case an input KATAKANA (square form of Japanese syllabary) input character string includes an unknown word of a signal character that is not contained in a dictionary. CONSTITUTION:A KANJI (Chinese character)-KANA (Japanese syllabary) document supplied from a keyboard 2 is stored in a document file 4 via a document input control part 5 and then divided in each matching process unit by a main control part 3 These divided documents are sent to a matching word identifying part 6 and matched with the words in a dictionary 7. The words in the dictionary 7 receive the information on the reading, accent, etc., and are converted into synthetic voices by a rule synthesizing part 9 to be outputted through a speaker 10. While the KATAKANA characters which are not contained in the dictionary 7 are sent to a KATAKANA processing part 8. Then an unknown KATAKANA word of a single character is included into its immediately preceding KATAKANA word for the second work identification. Thus the reading, accent, etc., of the KATAKANA word are sent to the part 9.
|
