Verbe Prediction methodes for the Korean language Disorders

초록

The language disorders hav difficulties in communication through the voice. so they have to be assisted by other devices. One of them is a sentence generation system which should be adaptable to ACC. It uses symbolizations, message and message prediction techniques. In this paper, we suggested the method which should apply the verb prediction for language disorders and confirmed its usefulness. Verb prediction is the methods which suggest the verb that persons frequently should use in one domain. The most useful verb is positioned at the top level in the pop-up menu. For the purpose, we used the neral network and simple Baysian Classifier. Each vocabulary was represented as a feature-vector according to semantics. When the user pushes the symbol, the right vocabulary searches the position in the state space. And the subject and object are combined and the verb are predicted. We could put the most frequently used verb on the top of the list. As a result, we could reduce the required time in the sentence generation comparing.