bilingual evaluation understudy

标签：combine   geo   bad   enc   evel   multi   wiki   modified   tip   

BLEU is designed to approximate human judgement at a corpus level, and performs badly if used to evaluate the quality of individual sentences.

https://en.wikipedia.org/wiki/BLEU

To produce a score for the whole corpus the modified precision scores for the segments are combined using the geometric meanmultiplied by a brevity penalty to prevent very short candidates from receiving too high a score.

bilingual evaluation understudy

标签：combine   geo   bad   enc   evel   multi   wiki   modified   tip   

原文地址：http://www.cnblogs.com/yuanjiangw/p/7553408.html