Construct chunk-level templates for improving rule-based machine translation

For improving the conventional rule-based machine translation by using translation templates, we propose a data-driven and semiautomatic method, which can semiautomatically extract translation templates at chunk level from the unannotated Chinese corpus. The method includes four steps: morphological analysis, chunk analysis, extract and refine. After extracting and preforming the preliminary templates, we manually add English translations for them and then get the ultimate templates, which are used in a template-based machine translation system. The experimental results show that the method is effective to improve the quality of machine translation, and that the template-based machine translation system outperforms the conventional rule-based machine translation system without templates. 1553-9105/Copyright © 2013 Binary Information Press.