Research on Automatic Error Correction Method in English Writing Based on Deep Neural Network

As one of the most widely used languages in the world, English plays a vital role in the communication between China and the world. However, grammar learning in English is a difficult and long process for English learners. Especially in English writing, English learners will inevitably make various grammatical writing errors. Therefore, it is extremely important to develop a model for correcting various writing errors in English writing. This can not only be used for automatic inspection and proofreading of English texts but also enable students to achieve the purpose of autonomous practice. This paper constructs an English writing error correction model and applies it to the actual system to realize automatic checking and correction of writing errors in English composition. This paper uses the deep learning model of Seq2Seq_Attention model and transformer model to eliminate deep-level errors. Statistical learning is combined with deep learning and adopted a model integration method. The output of each model is sent to the n-gram language model for scoring, and the highest score is selected as output.