Corpus-based correlational study of terms and quality in Business English writing

Abstract

One of the most important tasks in automated essay scoring (AES) is feature selection. Terms are indispensable in Business English (BE) writing. In order to analyze the possibility of involving terms in BE writing automated scoring feature set, the strength of correlations between terminological features and writing quality or scores is studied. A Business English term bank (BETB) was built based on a term dictionary. With BETB and a self-coded Python program, business terms and their categories in a BE writing corpus were identified and extracted. The analysis shows that, among ten categories of terms and total term numbers in BE writing, human resource terms and total term numbers have a moderate correlation with writing scores. This result means business terms, especially writing content related terms, should be covered in business AES feature set, which can improve the performance of AES systems and facilitate BE learners’ writing proficiency.