The purpose of this study is to establish the chinese-korean machine translation system. The study of the machine translation has received relatively little attention from korean scholarship in general, what was worse, that is much neglected branch of...
The purpose of this study is to establish the chinese-korean machine translation system. The study of the machine translation has received relatively little attention from korean scholarship in general, what was worse, that is much neglected branch of chinese linguistics in korea. So far. very little has been done in this direction. In China, this topic of the study has been studied extensively during the past five decades. The generation of sentence. till now. has mainly examined in syntactic dimension. That is to say, throughout these studies runs a common theme: the algorithm of machine translation system must be based on Chom`s rewriting rules. However, it cannot be satisfactorily explicated by syntax analysis. Because the sentence is not generated from the combinations between the syntactic categories. The structure of sentence is strong association with cooccurrence restrictions. Therefrore, we need more concrete and delicate the cooccurrence restrictions such as collocational restrictions. My approach in this paper is based on collocational restrictions and semantic informations. First, we pointed out the importance of punctuation in analyzing the structure of chinese sentence. Second. in considering language dependent phenomena. we would rather use morphologidcal alalysis than use word segmentation. This kind of processing may all the more possible because Chinese sentence has no word delimiter. Third Qi hu YAng`s (2000) study introduces the established researches on disambiguation. His approach is based on the principle of "preference". This principle seems plausible, yet it requires further examination. THe purpose of my paper is to an alternative approach which might present the method in solving the question. Four, we introduce some kind of method for establishing machine readable dictionary. we find out the method for dealing with equivocation. name and classify the collocational constructions. This study is only a first step. we believe it offers some insight into the problems. There are, however, a number of problems that remain to be explpred. The study concludes by mentioning some of the further problems raised by this approach to spreading activation model.