The Improvement And Implement Of An English Grammar Checking System Based On N-gram

English composition reflects the ability to write English, it’s an important part in English learning. In the learning process in the essay, there will be a very large number of writing exercises, but the number of English teacher is not enough to mark so many compositions. To solve the problem, it’s necessary to use the computer to aid to the composition marking. The syntax checking is an important part of essay marking. Qaunbin Wang have done the work on research of n-gram model, design of the checking algorithm, training of the corpus, implement of the grammar checking system and experiment on the effect of the checking system. The past system only take2-gram into account, and the checking algorithm is simple. This paper intends to improve the system.To improve the effect of the past checking system, in this paper improvements are achieved in three dimensions. First, the training of n-gram of POS tagger is added in training model, secondly the data in database is smoothed, thirdly the algorithm is optimized, the checking of POS tagger’s n-gram is added. In this paper the work is mainly on the design and implement of training system and inspection system. Work on the training system is mainly about the design and implement of the training algorithm, the redesign of the table in database, the choice of corpus. The work on inspection system mainly include the design and implement of the core algorithm and class.After the system implementation, the paper use the CLEC corpus to calculate the precision and recall rate of this inspection system. The experimental results show that the accuracy rate of the system increased by2%, while the recall rate of the system increased by10%. Compared with the system before, the improvement is obvious.