An Analysis Model of Potential Topics in English Essays Based on Semantic Space

Guimin Huang ； Xiaowei Zhang

《電腦學刊》 33卷1期 (2022/02) Pp. 151-164

https://doi.org/10.53106/199115992022023301014

摘要

With the reform of English examination in China in recent years, the automatic evaluation of English essays as subjective questions has always been the focus and difficulty of research. The existing automatic essay evaluation system (AEE) has obtained good feedback on the vocabulary, syntactic and other features of English essays, but there is still a problem of low accuracy in the analysis of potential topic in English essay. In order to solve this problem, this paper takes relational triples as the carrier to analyze the potential topics of English essays. By constructing the hierarchical topic trees hybrid semantic spaces to carry out topic clustering, distributed representation of topic relational triples and topic set extension in English essays. Then, based on the improved on-topic analysis algorithm in this paper, the paper analyzes the topic of English essay in multiple dimensions to obtain more abundant potential on-topic semantic information. The experiment results show that the proposed model can reduce the noise caused by non-topic words effectively, and improve the fine-grained topic semantic space in English essays, and the proposed model has better performance than the current methods of on-topic analysis in English esssays.

關鍵字

English essays ； topic analysis ； semantic spaces ； hLDA ； WP-nCRP