针对专利文本翻译中长句翻译的问题,提出了一种句子切分算法,即把长句切分成多个独立的小句后再进行翻译的方法.切分算法使用了概念层次网络(HNC)理论的语义特征来切分小句,切分算法和一个基于规则的基线翻译系统进行了融合,融合后的翻译系统的BLEU值达到0.189 8,比融合前的系统提高了30％.实验结果证明,提出的方法可以有效地改进专利翻译效果.%This paper presented a method using sentence segmentation to improve the performance of long patent sentence translation. In this method, segmented the long Chinese sentence into separated short sentences using some semantic features from the hierarchical network of concepts theory ( HNC theory). Integrated the segmentation algorithm with a rule-base MT system. After integrating, the BLEU value (0. 1898) of MT system was improved about 30%. The experiment results indicate the proposed method can improve the performance of patent sentence translation effectually.
展开▼
