Three Algorithms for Word-to-Phrase Machine Translation

摘要

Word-toȁ3;phrase machine translation (WPMT) is a model for lexical gap solving. Generally, lexical gap is an absent of word in a particular language. In a word-to-phrase model, words in source language could be replaced by target language phrases. This approach is very helpful for English- Vietnamese rule-based machine translation. The phrase replacement changes target sentence to unpredictable structure. This work develops three algorithms for WPMT: The insert algorithm applies phrase replacement, pruning algorithm simplifies target structure and moving algorithm keeps correct order of components in Vietnamese sentence.