Collecting Bilingual Technical Terms from Patent Families of Character-Segmented Chinese Sentences and Morpheme-Segmented Japanese Sentences

In manual translation of patent documents, a technical term bilingual lexicon is inevitable for a translator to efficiently translate patent documents. Dong et al. (2015) proposed a method of generating bilingual technical term lexicon from morpheme-segmented parallel patent sentences. The proposed method estimates Japanese-Chinese translation of technical terms using the phrase translation table of a statistical machine translation model. The procedure of generating bilingual technical term lexicon consists of the following four steps: (1) extracting Japanese technical terms from Japanese side of parallel patent sentences, (2) collecting all the sentences that contain the extracted Japanese term, (3) generating Chinese translation of the Japanese technical term referring to the phrase translation table of a statistical machine translation model, and (4) applying the Support Vector Machines (SVMs) to the task of identifying bilingual technical terms. In this paper, we segment the Chinese sentences into characters instead of segmenting them into morphemes as in Dong et al. (2015), and represent JapaneseChinese patent families in terms of character-segmented Chinese sentences and morphemesegmented Japanese sentences. Then, to those Japanese-Chinese patent families, we apply the framework (Dong et al., 2015) of identifying bilingual technical terms. As a result, we achieve the performance of over 90% precision with the condition of more than or equal to 60% recall.