Automatic pre-translation editing method and system for machine translation based on controlled language

The invention discloses an automatic pre-translation editing method and system for machine translation based on a controlled language, machine translation quality is improved by adopting a means of combining language rules and deep learning, a pre-translation editing neural network model is constructed, a to-be-translated original text corpus is processed, and the translation efficiency is improved. Error types of sentences with translation errors in an original text corpus are judged through the encoder, the sentences are rewritten through the decoder according to the error types of the sentences, and finally a final pre-translation result is obtained after correction is conducted through the copying mechanism module. According to the method, deep learning is carried out on the controlled language rule through the neural network model, the conditions of ambiguity, redundancy, logic chaos and the like of the original text can be remarkably reduced, and therefore the accuracy and fluency of machine translation