摘要 |
PROBLEM TO BE SOLVED: To easily detect whether outputted natural-language expressions are known as coocurrence or not. SOLUTION: For example, in the KANA(Japanese syllabary)-KANJI (Chinese character) conversion processing, an editing control part 102 checks whether a word group held in a word string candidate holding part 108 matches coocurrence information or not. When the word group in language information matches (or doesn't match) coocurrence information, the editing control part 102 adds coocurrence presence (or absence) information (for example, underline) indicating the result to a pertinent candidate character string to output it to a display part 103. Thus, it is easily detected by coocurrence presence (or absence) information given to the character string outputted to the display part 103 that expressions are judged to be known as coocurrence or are not stored as coocurrence. |