Several techniques have been developed to construct parallel corpus automatically. The most prominent system that generate parallel corpus from the World Wide Web is Structural Translation Recognition for Acquiring Natural Data (STRAND), developed by Resnik (1998, 1999). STRAND consists of three modules, candidate generation module, candidate evaluation module and candidate pair filtering module.
STRAND的构成
最新推荐文章于 2021-07-30 11:21:18 发布