|
To correct the article usage error in English texts,this paper proposes a novel approach based on classification for article error correction.However, there is only small quantity of labeled data available,while large quantity of data are available withoutannotations. To fully employ both types of data, we usethe balance-cascade algorithm to overcome the imbalanced data problem.Experiments were conducted on the NUS Corpus of Learner English and experimental results showed that the proposed method can achieve high precision rate. |
|
Keywords:Imbalanced data, Article error correction, Classification |
|