| 摘要 |
PURPOSE:To easily confirm the erroneous input of a KANJI (Chinese character) word whose reading is equal to that of a correct word but meaning cannot be understood at all, by detecting strings including a word consisting of one KANJI, into which a KANJI string consisting of two or more characters is divided, except the combination between numerals, the combination between a family name and a given name, or the like as the input error. CONSTITUTION:If Japanese-language sentence 'gonin no jido wa kenko de-' is inputted for example, this sentence is divided words 'go' 'nin' 'no', 'ji', 'do', 'wa', 'kenko', 'de'... by a sentence analyzing part 2. A meaningless KANJI string detecting part 5 selects strings including words consisting of one KANJI such as words 'go', 'nin', 'ji', and 'do', into which KANJI strings consisting of two or more characters are divided, and detects words 'ji' and 'do' except words 'go' and 'nin', which accord with the condition of a detection limiting part 8, as the input error, and the input error is reported to an operator. Thus, the erroneous input is surely detected when a KANJI word whose reading is equal to a correct word but meaning cannot be understood at all is erroneously inputted, and the erroneous input is easily confirmed to efficiently perform the sentence proofreading work.
|
