SPEECH SEMANTIC ANALYSIS PROGRAM, APPARATUS AND METHOD FOR IMPROVING COMPREHENSION ACCURACY OF CONTEXT SEMANTIC THROUGH EMOTION CLASSIFICATION

摘要

PROBLEM TO BE SOLVED: To provide a speech semantic analysis program or the like capable of improving a reliability of semantic comprehension by also taking emotion classifications of a user into account when comprehending context semantics from a speech of the user.SOLUTION: A computer is functioned as context semantic analysis means for outputting context reliability on a predetermined semantic classification basis regarding a text in a speech of a user; emotion reliability estimation means for estimating emotion reliability on a predetermined emotion classification basis from biological information of the user acquired by a device; and speech semantic complementation means for weighting the context reliability on the semantic classification basis with the emotion reliability on the emotion classification basis corresponding to the semantic classification, and selecting a semantic classification with the highest reliability. The context semantic analysis means outputs the context reliability of at least each of a positive semantic or a negative semantic, and the emotion reliability estimation means estimates emotion reliability of at least a positive emotion or a negative emotion and selects the positive semantic or the negative semantic with the higher reliability that is weighted by each of the emotion reliability.SELECTED DRAWING: Figure 1