| 摘要 |
PURPOSE:To split words even in a long sequence of characters by comparing a partial sequence of characters, obtained every time a sequence for an integral value preset from an initial character is erased by one character, with sequences of characters in a word dictionary. CONSTITUTION:The character length l0 of a KANA (Japanese syllabary) stored in a count buffer CB, i.e., the contents of a current counter LC are compared with a preset integral value W and while a sequence Cw1 of characters counted by W from the initial character of a character sequence C is extracted in a reference buffer 200, the W is set in a reference counter RC. Then, the character sequence Cw1 or a partial character sequence, obtained each time the character sequence Cw1 is erased by one character from the last tail character successively, is compared with character sequences in a word dictionary 600 sequentially, thereby storing a split-word buffer 300 with the longest character sequence among coincident character sequences as the 1st split word Cd1. Then, the remaining character sequence CR1 obtained erasing the split word Cd1 from the character sequence C is re-stored in the CB and split similarly and successively to generate the 2nd, 3rd-(n)-th split words. |
