Reading level assessment using support vector machines and statistical language models
ACL '05 Proceedings of the 43rd Annual Meeting on Association for Computational Linguistics
Hi-index | 0.00 |
We propose a reading time model for learners of English as a foreign language (EFL) that is based on a learner's reading proficiency and the linguistic properties of sentences. Reading proficiency here refers to a learner's reading score on the Test of English for International Communications (TOEIC), and the linguistic properties are the lexical, syntactic and discourse complexities of a sentence. We used natural language processing technology to automatically extract these linguistic properties, and developed a model using multiple regression analysis as a learning algorithm in combining the learner's proficiency and linguistic properties. Experimental results showed that our reading time model predicted sentence-reading time with a 22.9% error rate, which is lower than the models constructed based on linguistic properties proposed in previous studies.