Abstract
An NLP algorithm was developed to assess question quality to inform feedback on questions generated by students within iSTART (an intelligent tutoring system that teaches reading strategies). A corpus of 4575 questions was coded using a four-level taxonomy. NLP indices were calculated for each question and machine learning was used to predict question quality. NLP indices related to lexical sophistication modestly predicted question type. Accuracies improved when predicting two levels (shallow versus deep).
Author supplied keywords
Cite
CITATION STYLE
Kopp, K. J., Johnson, A. M., Crossley, S. A., & McNamara, D. S. (2017). Assessing question quality using NLP. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10331 LNAI, pp. 523–527). Springer Verlag. https://doi.org/10.1007/978-3-319-61425-0_55
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
