Web-based automatic writing assessment and analysis system for PSLE Chinese composition

• Main Author: Wan, Liuyang
• Other Authors: Hui Siu Cheung
• Format: Final Year Project
• Language: English
• Published: 2014
• Subjects: DRNTU::Science
• Online Access: http://hdl.handle.net/10356/61976
• Institution: Nanyang Technological University

Various techniques for automatic grading English compositions are available for production use. In contrast, no known and stable systems have done similar work to the Chinese language. This project identifies the difficulties in processing Chinese, studies existing solutions for automatic English assessment, and adapts the techniques to grade Chinese composition in the context of PSLE. Specifically, this project builds a web-based system to aid the question browsing and searching experience and to allow the users to receive instant feedbacks on their writing skills. Grading metrics include syntactic complexity, grammar rules, and semantic analysis. The author further proposed some new approaches: the use of Chengyu, mastery of punctuation, use of writing patterns, word count. Finally these metrics were evaluated and recommendations were given.