Practice and Analysis in Evaluation of Machine Translation
Cai Jing-qiu · Jisuanji yingyong yanjiu · 2004
Machine Translation is one of the most challenging problems in artificial intelligence and recently its evaluation gets more and more attention.In order to assess the translation quality and to pinpoint the problems in current Machine Translation System,this paper presents a method of using a large-scale corpus as a testing set by which we have semi-automatically evaluated six Machine Translation Systems.Finally we have analysed the problems of morphology and syntax in Machine Translation Systems,and have proved the objectivity of our evaluation result by comparing with automatic evaluation of NIST.