| 摘要 |
PURPOSE:To more effectively execute rule conversion processing without increasing the memory capacity of a rule dictionary by applying the same information to plural words belonging to the same category and describing the words in a rule item. CONSTITUTION:A 'KANA' (Japanese syllabary) reading character string is inputted from an input means 1 and 'KANA'/'KANJI' (Chinese character) converting means 2 executes the 'KANA'/'KANJI' conversion of the inputted character string based upon a reference dictionary 3 storing notation corresponding to the reading of each word and information expressing the category of the 'KANA'/'KANJI' conversion result is stored in a conversion result storing means 5 and a rule stored in a rule dictionary 6 is compared with the contents stored in the means 5 by a rule retrieving means 7 to retrieve the existence of a role having a word belonging to the same category and coincidence of a pattern for words other than the word. At the time of retrieving a coincident rule, the contents of the means 5 are rewritten by a rule rewriting means 8 in accordance with a retrieved rule rewriting command and a 'KANA'/'KANJI' character string obtained by rewriting the contents of the means 5 is outputted. |
