Automatic English grammar error checking and recommendation

• Main Author: Min, Arkar
• Other Authors: Hui Siu Cheung
• Format: Final Year Project
• Language: English
• Published: 2019
• Subjects: Engineering::Computer science and engineering
• Online Access: http://hdl.handle.net/10356/78971
• Institution: Nanyang Technological University

Traditionally, English grammatical error checking is done by English language professionals. However, due to the limited availability of those professionals and the increasing number of English learners, the need for automatic grammatical error-checking tools, which can be used anywhere and anytime has risen. This project built an automatic English Grammatical Error Checking system using neural machine translation (NMT). The NMT model was trained using the Transformer model architecture and Lang-8 and NUCLE datasets. This training process was further improved using source-word dropout, domain adaption and tied embedding techniques. The benchmark performance of the system was evaluated on JFLEG corpus using the BLEU score and this system achieved the score of “52.4”. The performance of the model on different types of grammatical errors was also evaluated. In addition, the project also included English Grammatical Error Explanation using rule-based checking. An intuitive and user-friendly web prototype system was developed to demonstrate the working of automatic grammatical error checking, correction, and explanation.