Research on intelligent machine translation error correction system based on data mining and language features
To improve the quality of machine translation and promote international communication.Based on data mining and lan-guage features,the intelligent machine translation error correction system is constructed.This paper studies the combination of multi-ple confidence features and the classification of translation errors using maximum entropy classifier.Then the translation is corrected based on the paraphrase extraction method to improve the quality of machine translation.The experiment shows that after the interven-tion of the error correction system,the average ELEU value of machine translation is 96.83％,which is 14.47％higher than that be-fore the improvement.The data shows that the error correction system can effectively identify and correct errors in machine transla-tion,thus improving the quality of translation,as a strong support for international communication.
MT data miningmaximum entropy classificationretelling extraction
万方数据
维普
