Web、语料库与双语平行语料库的建设

对Web和语料库以及多语语料库的关系进行辨析,针对Web上丰富的各类电子文本,从语言工程角度出发,提出"分步骤、按领域"建设大规模双语平行语料库的思路,即选定领域专一、语言可靠、格式规范的文本,逐次建设特定领域的语料库,最后汇总成高质量、大规模、全领域的"高大全"式双语平行语料库。同时,围绕一个实例介绍如何利用Web资源建设特定领域双语平行语料库。

Web, Corpus and the Building of Bilingual Parallel Corpora

There are different understandings of Web as corpus. We try to explore the relations between Web, corpus and bilingual parallel corpora. Inspired by the rich electronic texts available on World Wide Web, and the strategy of sublanguage in language engineering, we propose a solution to building a large-scale bilingual parallel corpus, by accumulating homogeneous documents in different domains. The large amount of texts with high quality on a restricted domain collected at each step eventually constitutes a massive general-purpose balanced data warehouse. An example is elaborated to show how to construct a domain-specific bilingual parallel corpus from the Web.