Detecting Errors in English Article Usage with a Maximum Entropy Classifier Trained on a Large, Diverse Corpus
Na-Rae Han, Martin S. Chodorow, Claudia Leacock · 2004
One of the most difficult challenges faced by non-native speakers of English is mastering the system of English articles.We trained a maximum entropy classifier to select among a/an, the, or zero article for noun phrases, based on a set of features extracted from the local context of each.When the classifier was trained on 6 million noun phrases, its performance was correct about 88% of the time.We also used the classifier to detect article errors in the TOEFL essays of native speakers of Chinese, Japanese, and Russian.Agreement with human annotators was about 88% (kappa = 0.36).Many of the disagreements were due to the classifier s lack of discourse information.Performance rose to 94% agreement (kappa = 0.47) when the system accepted noun phrases as correct in cases where its own confidence was low.