| 摘要 |
PROBLEM TO BE SOLVED: To provide a voice processing device capable of exactly analysing input voice without accompanying a troublesome work and outputting by inputting voice to convert into character data, and analysing the character data for every speaker and outputting it. SOLUTION: The RAM 103 of a POS terminal is provided with a voice recognizing dictionary for every speaker housing part housing a voice recognizing dictionary to recognize input voice, and a knowledge dictionary for every speaker housing part housing a knowledge dictionary to store information such as a part of speech of a word, concatenation, changed form of a word, concepts used for analysis of natural word, and provided corresponding to them. Voice recognition circuits 114, 115 recognize the input voice referring to the voice recognizing dictionary of the voice recognizing dictionary for every speaker housing part, and after converting into character data, voice data and character data are recorded as voice information, and the voice data up to completing voice recognition process is controlled as conversion data. The character data is read out, analysed by the use of the corresponding knowledge dictionary of the knowledge dictionary for every speaker housing part, and the analysed result is output. |
