| 摘要 |
PURPOSE:To reduce a dispersion of a selection executed by a user, by executing the maximum evaluation by a length of reading of a word, use frequency information and a connection weight, and selecting a prescribed word based on output sequence information of the word. CONSTITUTION:A directionary retrieving device 3 stores reading and KANJI (Chinese character) information, information for showing the frequency of a part of speech, information of a length of reading of a word, and information of a part of speech of the word, in a word buffer 4. A connection weight mem ory 5 studies whether a part of speech of a preceding word can be connected to the word concerned or not, and inputs its result to an evaluating and operat ing circuit 6. The evaluating and operating circuit 6 calculates the maximum evaluation value, based on a connection weight between inputted parts of speech, a length of reading of the word which has been fetched from the word buffer 4, and frequency information of the part of speech. Based on its result, by a sorting circuit 8, the same sound word is discriminated by a sequence learning and outputted.
|
