NEW MMS DATA-INPUT METHOD BASED ON THE CONCEPT OF VAGUENESS AND SUBJECTIVE ENTROPY.

Human 'ambiguity' is one of the most important factors in (conversational) man-machine-systems (MMS's). Two problems in MMS's are dealt with; one is how to express ambiguous data given by human beings, the other what is the optimal number of states in answering questions given by machines in the case that only ambiguous data are available. A decision making model is proposed. A new MMS data-input method is proposed, where 5-valued membership information is given by human beings with an appropriate use of 3-valued vagueness information. Availability of this method is also confirmed by a simulation experiment in human character/voice pattern recognition process.