An Automatic English Composition scoring model based on neural network algorithm

In this paper, an Automatic English Composition scoring (AECS) model based on neural network algorithm is constructed by extracting the lexical feature, syntactic feature and readability features which reflect the content writing quality and determining these features´ weight in composition scoring. The model uses training data to train the neural network and eventually it obtains the neural networks indicating the relationship of these features which can be used to predict the English compositions´ final scores. Through an objective comparison of the scores predicted by AECS and experienced teachers, we know that the AECS model we proposed can well reflect the level of students´ writing.