| 摘要 |
PROBLEM TO BE SOLVED: To collate a KANA character string with a correct word without performing separation and integration of voiced sound symbol or a semivoiced sound symbol attached to KANA by mutually shifting the character codes of a unvoiced consonant character and a voiced consonant or a p-sound sign character to the same state. SOLUTION: A word collating part 1 comprises a set of states Si and each state Si has a pointer that points information in a word dictionary part 2 and a transition possible character table. For instance, when a word of the dictionary 2 is 'miyazaki', in the character 'mi', the correspondence relation of the character code of the 'mi' and a pointer that points a state S2 is stored in the transition possible character table in an initial state S1, and in the state S2, the next character 'ya' is paid attention to. Thus, in the process of repeating a transition operation, the transition possible character table stores correspondence relation about voiced consonant characters 'za' and 'gi' which add a voiced sound symbol to unvoiced consonant characters 'sa' and 'ki' which he voiced sound symbol can be added to so that they may be shifted to the same states S4 and S5 and stores the correspondence relation about a voiced sound symbol character so that it may shift to the state S5 itself that is a transition destination. |
