基于二元语篇信息序列的译文抄袭鉴定方法研究

A Study on Translation Plagiarism Detection Methods based on Discourse Information Bigrams

由于现有研究中缺乏针对改变句法顺序这一译文抄袭策略的鉴定方法,因此,本文运用语篇信息理论,利用判别分析,论证了二元信息单位序列在真正译者译文中分布一致性较好,而在抄袭者译文中的分布则差异较大。此外,本文还通过构造二元信息成分序列,再次验证了其用于鉴别同义词替换这种抄袭策略的有效性。最后,针对个人言语特征理论依据不足的问题,本文以语篇信息为视角,讨论了二元语篇信息序列作为个人言语特征的优势所在。

In the existing literature of translation plagiarism detection,little research has been done to deal with the change of syntactic order.On this account,the researcher employs Discourse Information Theory,conducts discriminant function analysis and finds that the distribution of information unit bigrams exhibits a fairly good homogeneity in the true translator s texts but an obvious heterogeneity in the plagiarizer s texts.In addition to this,the effectiveness of information element bigrams is proved again in detecting synonym substitution.At last,as idiolect is lack of solid theoretical foundation,the advantages of treating discourse information bigrams as idiolect are also discussed in this article.