计算机应用研究2017，Vol.34Issue(12)：3538-3542,3546,6.DOI:10.3969/j.issn.1001-3695.2017.12.004
基于语义语法分析的中文语句困惑度评价
Evaluating perplexity of Chinese sentences based on grammar & semantics analysis
摘要
Abstract
Methods currently used to evaluate the translation quality of machine translation system are BLEU introduced by IBM,TER,METEOR.These methods do evaluation according to the recurrence rate,the edit distance and the linguistic knowledge,respectively.Although they are superior to other methods,they have some defects in judging the perplexity of Chinese sentences.This paper proposed a method for evaluating the perplexity of Chinese sentences.Specifically,it analyzed the grammar and the semantics together for Chinese sentences and their trunks,based on the parsing of dependency grammar.In the comparison experiments,this method outperforms the mainstream evaluating ones,with accuracy 4 percent higher than improved weighting BLEU method.关键词
困惑度/病句/语法/语义/机器翻译Key words
perplexity/wrong sentences/grammar/semantics/machine translation分类
信息技术与安全科学引用本文复制引用
何天文,王红..基于语义语法分析的中文语句困惑度评价[J].计算机应用研究,2017,34(12):3538-3542,3546,6.基金项目
国家自然科学基金资助项目(61672329,61373149,61472233,61572300,81273704) （61672329,61373149,61472233,61572300,81273704）
山东省科技计划资助项目(2014GGX101026) （2014GGX101026）
山东省教育科学规划资助项目(ZK1437B010) （ZK1437B010）
山东省泰山学者基金资助项目(TSHW201502038,20110819) （TSHW201502038,20110819）
山东省精品课程资助项目(2012BK294,2013BK399,2013BK402) （2012BK294,2013BK399,2013BK402）
