| 摘要 |
PURPOSE:To convert intended KANJI (Chinese character) without correcting a character string even if the wrong character string is inputted by obtaining the degree of coincidence by subtracting the coincidence evaluation information of a non-coincident character from the maximum value of the coincidence evaluation information. CONSTITUTION:This KANJI converter consists of a character input part 1, a character analyzing part 2, a retrieving part 3, a dictionary part 4, a coincidence detecting part 5, and a converting part 6. Then, the coincidence evaluation information is attached to each character of an index set in the dictionary part 4, and the maximum value of the coincidence evaluation information in that index is determined, and it is discriminated whether an inputted KANA (Japanese syllabary) character string decomposed into clauses and the KANA character string of the index set in the dictionary part 4 coincide with each other or not for every one character, and only at the time of non-coincidence, the coincidence evaluation information of the non-coincident character is substrated from the maximum value of the coincidence evaluation information. Then, when the degree of coincidence at the stage of the end of coincidence retrieval is over a prescribed level compared with the maximum value of the coincidence evaluation information, the KANJI of a candidate corresponding to that index is selected. |
