This article describes the experiments we performed during our participation in the HOO Challenge. We present the adaption we made on two systems, mainly designing new grammatical rules and completing a lexicon. We focused our work on some of the most common errors in the corpus: missing punctuation and inaccurate prepositions. Our best experiment achieved a 0.1097 detection score, a 0.0820 recognition score, and a 0.0557 correction score on the test corpus. 1

We present a novel beam-search decoder for grammatical error correction. The decoder iteratively generates new hypothesis corrections from current hypotheses and scores them based on features of grammatical correctness and fluency. These features include scores from discriminative classifiers for specific error categories, such as articles and prepositions. Unlike all previous approaches, our method is able to perform correction of whole sentences with multiple and interacting errors while still taking advantage of powerful existing classifier approaches. Our decoder achieves an F1 correction score significantly higher than all previous published scores on the Helping Our Own (HOO) shared task data set. 1