Automatic Classroom Question Classification Based on Bloom's Taxonomy

Abstract

Asking questions is usually used by teachers to guide students to think and to interact with students. Bloom's Taxonomy has been used widely in the educational field to assess students’ intellectual abilities and skills. However, most questions lack a scientific basis and design; too many low-level questions constrain students’ thought. Some works classify these questions manually, which is inefficient. To improve the efficiency and provide implications for teachers to design curriculums, this study utilized machine learning to automatically classify teachers’ questions by building up keywords and extracting TF-IDF (Term frequency inverse document frequency) features. The result showed that keywords are significant in classifying questions, and we obtained an accuracy of 86.0%.