Automatic grammar error correction model in English writing teaching

Abstract

This paper developed a automatic grammar error correction (GEC) model based on the sequence-to-sequence (seq2seq) model. It adopted a dual-encoding structure composing of a syntactic encoder and a semantic encoder, and introduced a hybrid attention mechanism in the decoder. Finally, the results were output through the softmax layer. Experimental results on two general GEC test sets, CoNLL-2014 and JFLEG, showed that this model outperformed the existing comparison models in terms of precision, recall, and comprehensive evaluation indicator. In the error-correction test of actual English compositions written by students, the model successfully corrected more than half of the grammar errors marked manually. The results show that this model has good GEC performance and is suitable for English writing teaching.