Syntactic analysis and hierarchical phrase model based machine translation system and method

A syntactic analysis and hierarchical phrase model based machine translation method, the method comprising:

• obtaining a syntactic based noncontiguous phrase rule set through a phrase alignment table obtained from a bilingual aligned text and a syntactic annotated corpus obtained from an annotated corpus;
• receiving an annotated corpus and the bilingual aligned text, extracting useful language knowledge and probability information from the annotated corpus, annotating bilingual or monolingual in bilingual aligned text with respect to gender and syntax based on the abstracted useful language knowledge and probability information, and generating the syntactic annotated corpus;
• and
• outputting a translation result with respect to input sentences through the syntactic based noncontiguous phrase rule set.

A machine translation method comprising:

• receiving a bilingual aligned text and an annotated corpus;
• generating a bilingual aligned text based on the phrase to be translated;
• generating syntactic annotated corpus based on the annotated corpus and the bilingual aligned text by extracting useful language knowledge and probability information from the annotated corpus, annotating bilingual or monolingual in bilingual aligned text with respect to gender and syntax based on the abstracted useful language knowledge and probability information, and generating the syntactic annotated corpus;
• generating a phrase alignment table based on the bilingual aligned text;
• generating a syntactic based noncontiguous phrase rule set based on the syntactic annotated corpus and the phrase alignment table;
• machine translating an input sentence into a target language based on at least one of the phrase alignment table and the syntactic based noncontiguous phrase rule set;
• evaluating results of the machine translation based on an evaluation model;
• and
• outputting, as a translated sentence, a result of the evaluation having a highest score among the evaluated results.