Dynamical verb prediction for korean language disorders in restricted domain

초록

In this papaer. we proposed the augmentative communication system that generated well-formed sentences in Korean language. In generating sentence, symbols and symbol sequence are used to represent lots of vocabularies easily within the restricted domain. There is a limitation of requiring learning and quess in such a dynamic and static system. The system includes parts of display, control and contents. The display part included input words and sentences. The comtrpl part includes the place domain and fuction of verbs, adjective, adverb and can include the core vocbulary. In the contents part, the icons for nouns of actor part and objects are located at left side and icons for verbs are language disorders to be able to communication with normal people naturally. When we compar relation each sentences with methods, we can see a more enhancement than the other method in verb prediction except uncommon sentence in the result of experiment. Therefore, we think that the prediction is useful.