Rule-based Phrase Identification in College English Writing

Author

Ge, Shi-Li ; Song, Rou

Author_Institution

Nat. Key Res. Center for Linguistics, Guangdong Univ. of Foreign Studies, Guangzhou, China

fYear

2010

fDate

13-15 Dec. 2010

Firstpage

240

Lastpage

243

Abstract

In order to support automated essay scoring and feedback offering in college English writing, phrases used in compositions, including correct usage and wrong usage, should be identified accurately. With the help of composition analysis and teaching experience, more than 1300 phrases frequently used in college English writing were picked out. A rule base of relevant phrase identification rules were constructed, including more than 1300 correct usage rules and 200 wrong usage rules. For illustration, a set of 660 college students´ compositions was adopted to test the application of the rules. Experimental results show that the precision and recall of phrase identification reach 96.28% and 82.99%, respectively, which means that the rule base is applicable for automated essay scoring and feedback offering in Chinese college English writing.

Keywords

educational institutions; feedback; identification; knowledge based systems; natural language processing; teaching; text analysis; Chinese college English writing; automated essay scoring; composition analysis; rule-based phrase identification; teaching experience; Artificial neural networks; Educational institutions; Grammar; Pragmatics; Tagging;

fLanguage

English

Publisher

ieee

Conference_Titel

Genetic and Evolutionary Computing (ICGEC), 2010 Fourth International Conference on

Conference_Location

Shenzhen

Print_ISBN

978-1-4244-8891-9

Electronic_ISBN

978-0-7695-4281-2

Type

conf

DOI

10.1109/ICGEC.2010.66

Filename

5715414