Recognition and Transformation of Normal Sub-Sentence in Chinese-English Patent Machine Translation

Normal sub-sentence is a common language phenomenon. The differences between Chinese and English normal sub-sentence make a transformation processing necessary in machine translation. This paper proposed a method to recognize and transform normal sub-sentences to improve their translation performance in patent machine translation. First, we created a dictionary based on HNC theory to recognize the normal sub-sentence. Then, we compared Chinese normal sub-sentences to English ones and summarized the transformation rules based on the comparison. At last, we applied the method to an existing machine translation to test the processing effect. The experiment results show that the method can effectively recognize and transform normal sub-sentences. The Blue of the translation is raised by about 7%.