An Analysis of the Evaluation of the Translation Quality of Neural Machine Translation Application Systems

Neural machine translation (NMT) is applied to generate a more reliable and accurate practice as the most cutting-edge technology. In recent years, NMT has achieved gratifying results. However, main obstacle for market-oriented application systems appears suffer from weak quality that fails meet users’ needs. This paper focuses on of political documents implements six dominant in market evaluate their quality. The evaluation process further employs both BLEU NIST technical algorithms re-verifies results with manual method called “Score Ranking System” compare performances NMTs Chinese-English translations documents. Through diagnosis problems errors NMTs, eventually proposes “Cue Lexicon+” model remedy prominent problems. Besides, “NMT+ Lexicon Intelligent Translation Assistant” soft developed integrated into improve quality, providing reference research basis increase performance update systems.