A new method of Figurative Rhetoric Recognition based on Automated Essay Scoring of the Oversea Chinese Students’ Instructional Composition Corpus

Abstract

With the rapid development of modern science and technology and the rapid increase of new scientific and technological knowledge and information, mankind has entered the so-called era of information explosion. In terms of human reading experience, information explosion leads to the change of reading mode, which has brought negative effects and potential crises to human society. The automated essay scoring has great educational and commercial values, and provides a cost-effective and consistent alternative to human marking. The figurative rhetoric is important to the ideological content and artistic quality of the essay. Based on large-scale language text data, various natural language processing technologies have made great progress in the fields of text classification, speech recognition and so on. This paper proposed the experimental methods such as CNN, RNN, Transformer, Fast Text and Bert base to recognize the figurative rhetoric based on the oversea Chinese students' composition corpus. We find that the Bert model has achieved an accuracy of 80.93% in the identification of figurative rhetoric sentences in oversea Chinese student compositions. The experimental methods in this paper are feasible, and can promote the improvement of the automated essay scoring technology.