Automatic writing assessment for ESL students

• Main Author: Gong, Mingyi
• Other Authors: Hui Siu Cheung
• Format: Final Year Project
• Language: English
• Published: 2016
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/66641
• Institution: Nanyang Technological University

In this project, an automatic writing assessment system is designed and developed. It focuses on analysis and assessment on composition questions of CET (College English Test) which is a national ESL (English as a second language) test held in People's Republic of China. The target candidates of CET contain both undergraduate and postgraduate non-English-specialized students in the universities. There are two bands of CET, CET4 and CET6. Results of these two tests represent the English proficiency of students, and are highly concerned by companies and universities if the students would like to be hired or engage in advanced studies. And among all CET questions, skill on composition writing is the most difficult for students to improve. This project aims to build a computerized automatic English compositions assessment system and provide non-English-specialized students with a web-based environment to browse and practice CET composition questions. Multiple assessment tools and techniques are studied and implemented in the system, in order to generate both short-term and long-term writing feedbacks based on five metrics, syntactic complexity, lexical richness, semantic correlation, grammar and spelling correctness. Various experiments have been conducted and performance of the system is evaluated. Suggestions on future improvements are also proposed.