Applying machine learning to analyse teachers' instructional questions
Anwar Ali Yahya, Mohammad Said El Bashir · International Journal of Advanced Intelligence Paradigms · 2014
This paper introduces machine learning approaches to a new application in the field of education. More specifically, it explores the effectiveness of three machine learning approaches, namely, k-nearest neighbours, naïve Bayes, and support vector machines with term frequency as term selection approach, on the task of evaluating teaching effectiveness by classifying teachers’ classroom questions into different cognitive levels identified in Bloom’s taxonomy. In doing so, a dataset of questions has been collected and annotated manually with Bloom’s cognitive levels. Several steps of pre-processing have been applied to convert these questions into a representation suitable for machine learning approaches. Using the dataset, the performance of the machine learning approaches and the traditional rule-based approach have been evaluated. The obtained results lead to several conclusions: First, machine learning approaches have a superior performance over rule-based approach. Second, the term frequency as a term selection approach plays a crucial rule in the performance of machine learning approaches. Third, SVM shows a superior performance over k-nearest neighbour and naïve Bayes which shows a comparable performance in term of F-measure and accuracy. Finally, machine learning approaches show different levels of sensitivity to the number of terms used in question representation.