Chuyển đổi cây cú pháp cụm danh từ tiếng Anh theo tiếng Việt

Abstract

Transferring syntax trees is one of the key tasks of machine translation systems. To transfer syntax  trees,  they  can  be  performed  by  different models  of  the  statistical  translation  method,  rulebased translation method, or a hybrid of statisticsand rule-based  translation  method  [9].  In  this  paper  wepresent  a  combination  of  bilingual  corpus  and knowledge-based method, which transfers syntax trees of  English  Base  Noun  Phrases  via  the  Vietnamese syntax  from  a  pair  of  bilingual  sentences  to  identify anchor points. Our combination technique can help to invert  word  order  in  noun  phrases  of  the  source language to suit those of target language and improve the  performance  of  miss-alignment,  null-alignment, overlap  and  conflict  projection  of  the  existing methods.  The  proposed  technique  can  be  easily applied to other language pairs. Experiment on pairs of  sentences  in  the  English-Vietnamese  bilingual corpus  showed  that  our  proposed  method  is satisfactory.