摘要 |
PURPOSE: A method for setting units of similar features in a continuous voice recognition system is provided to increase the capability of the voice recognition system. CONSTITUTION: A method for setting units of similar features in a continuous voice recognition system includes a numeric counter(nnc), time common noun(nct), amount counter. The method uses KTS(Korean tagging system) feature symbols. a state noun(ncps) and an adjective oriented tail(xsm) is classified into a status adjective(paa). The status adjective(paa) and another adjective tail(efm) are classified into a status adjective(mma). Words with an adverb oriented tail(xsa) is classified into a common adverb(mas). A common calling tail is combined in the connection tail(jcj). The previous tail(ep) is classified into a word tail. Information representing status is classified into an auxiliary tail. |