摘要 |
PROBLEM TO BE SOLVED: To provide a method and apparatus for classifying spoken language into at least one of a plurality of categories. SOLUTION: In this method and apparatus, the spoken language is converted into a text and confidence score is provided to one word or a plurality of words at the time of conversion. The spoken language is classified into at least one category on the basis of (1) the degree of approximation among words at the time of the conversion of the spoken language and a word of at least one category and (2) the confidence score. For example, the degree of approximation is made the size of the cosine degree of similarity among the query vector display of the spoken language and the plurality of categories. The scores are generated to the plurality of categories respectively with optional selection and these scores are used for classifying the spoken language into at least one category. For example, the confidence score of a word constituted of a plurality of words can be calculated as the geometric means of confidence scores of respective words of the word constituted of the plurality of words. COPYRIGHT: (C)2006,JPO&NCIPI
|