Chinese-English Parallel Corpus Construction and its Application

Chinese-English Parallel Corpus Construction and its Application CHANG Baobao Institute of Computational Linguistics Peking University Beijing, P.R.China chbb@pku.edu.cn Abstract Chinese-English parallel corpora are key resources for Chinese-English cross-language information processing, Chinese-English bilingual lexicography, Chinese-English language research and teaching. But so far large-scale Chinese-English corpus is still unavailable yet, given the difficulties and the intensive labours required. In this paper, our work towards building a large-scale Chinese-English parallel corpus is presented. We elaborate on the collection, annotation and mark-up of the parallel Chinese-English texts and the workflow that we used to construct the corpus. In addition, we also present our work toward building tools for constructing and using the corpus easily for different purposes. Among these tools, a parallel concordance tool developed by us is examined in detail. Several applications of the corpus being conducted are also introduced briefly in the paper. 1 Introduction Corpora are no doubt key resources for language information processing, language research and teaching and lexicography. Many large-scale monolingual corpora have been available in the world, especially with English corpora. Efforts to build Chinese corpora started around 1990s in Mainland China. Substantial progress has been made so far. There are several large-scale Chinese corpora available now, such as the People's Daily corpus developed by Peking University. However, Compared with the big progress in building monolingual corpus, there are not so many practical parallel corpora available, especially parallel corpora with Chinese involved, which hinders their use in cross-language information processing, language teaching and research and bilingual dictionary compilation. In order to give substantial support to research in related fields, we started compiling a Chinese-English parallel corpus in 2001. So far over 10 million character of Chinese texts of different types and corresponding English texts has been collected and included into the corpus. The corpus has been automatically aligned and verified manually at sentence level and the Chinese part also has been automatically segmented and POS tagged. Tentative researches using this corpus in Machine Aided Translation, bilingual dictionary compilation and other related field has been also conducted. We will present our work towards building and using the Chinese-English corpora in this paper. 2 The construction of the Chinese-English corpus 2.1 The collection of the parallel texts There is a wide literature on corpus design principles; one of them is that a general corpus should be made balanced. Which means a corpus shall contain texts of different domain and different genres in reasonable proportions; the corpus thus can be a reasonable reflection of the language use. However, when we decided to construct the Chinese-English corpus, we found it's not easy to construct a perfectly balanced Chinese-English corpus. That's because there are not so many electronic Chinese-English bilingual texts available. So we decide to collect Chinese-English bilingual texts as many as we can, as PACLIC 18, December 8th-10th, 2004, Waseda University, Tokyo - 283 - long as the texts are of good quality. For the same reason, we decided not to use any sampling techniques; the full texts are included in …

Content truncated to 500 words