| 摘要 |
PROBLEM TO BE SOLVED: To enhance the convenience of using customers by providing a menu selection function by a voice input capable of highly accurate voice recognition processing by a learning function utilizing a network and a function capable of making the using customer reach desired information without being conscious of the structure of a menu. SOLUTION: The speech of a using customer is frequency-analyzed and converted into the parameter of phoneme data or the like, and based on the parameter of the phoneme data or the like, a voice recognition processing is performed in order of phoneme recognition processing, word recognition processing and sentence recognition processing. As a word recognition dictionary at the time, a command voice table 413B and voice tables (contents voice table 413C, maker voice table 413D, genre voice table 413E and field voice table 413F) relating to the items of guidance display are referred to. In the case that a word inside the voice table relating to the item of the guidance display is recognized, information relating to the link state of the items set to name masters corresponding to them is referred to, a menu system is constructed and shifted to guidance output processing based on a guidance sentence data table 413L. |
