| 摘要 |
PURPOSE:To extend the range of success in coocurrence to an unregistered word, and to improve the efficiency of conversion by judging the meaning of the corrected unregistered word by referring to a coocurrence dictionary and adding a corresponding semantic code to it when a user corrects the unregistered word of a dictionary. CONSTITUTION:This converter is constituted of an input device 1, a central control device 2 such as a CPU, etc., for controlling a whole device, a RAM 3 and a ROM 4 as a storage means, and an output device 7 such as a CRT, etc., for displaying the contents of an output buffer, etc. Then, a character string, etc., is inputted from the input device 1, and is converted by a KANA/KANJI converting part 51 by referring to a basic dictionary 61 to store the reading, the description, and part of speech information of the word and the semantic code to classify the word from the viewpoint of the meaning. At that time, when there is the unregistered word in the inputted character string, and the user corrects a converted result part containing this unregistered word, the meaning of this unregistered word is judged by a semantic judging part 53 by referring to the coocurrence dictionary 65 to store the semantic code and the word in coocurrent relation to this code as making them a pair, and the corresponding semantic code is added by a semantic code adding part 55. |
