Rule Reengineering Approaches Using SC-Grammar Analysis

摘要

Based on SC-Grammar with which we developed the intelligent machine translation(MT) system IMT-EC/863, this paper analyzes rule reengineering approaches in detail, taking advantages of the powerful predicative ability of SC-analysis. A lot of rule reengineering approaches, which are applicable in our existing machine translation systems, are presented. A very important thought of machine translation has been put forward, which is to combine some of corpus analysis methods and related achievements in the so-called rule-based machine translation in order to decrease the height of syntactic and semantic analysis tree remarkable. With such multistrategy approaches, machine translation process can be speeded up a great deal, enhancing the translation quality remarkably.