PNS2E: A POS-Based Grammatical Error Correction Model with Pre-Correction and Post-Selection

English grammatical error correction has become a hot topic in natural language processing domain in recent years. Since CoNLL-2014 Shared Task, Scholars have proposed new architectures, designed optimization algorithm, and implemented models whose performance were greatly enhanced. In this paper, we introduce a novel GEC model called PNS2E. By introducing a pre-correction checker and a post-selection language model, a better evaluation result than the baseline on the test set of the CoNLL-2014 Shared Task is achieved.