المستخلص: |
In quality assurance, Bloom's taxonomy can be used to automatically classify educational goals, objectives, learning outcomes and questions. Some educational organizations such as accreditation bodies are in need to check the correctness of the classification of exam questions according to Bloom’s Cognitive Levels and they might be faced with an enormous number of questions which would be difficult to check manually. Therefore, the usage of automatic classification of questions based on Bloom's taxonomy is highly needed. This paper aims to test and compare different machine learning methods (Naive bayes, support vector machine, logistic regression, and decision trees) to automatically classify exam questions based on the cognitive levels of Bloom's taxonomy. The features used in the classification were based on linguistically-motivated features which are the bag of words, part of speech (POS), and n-grams. A database contains 600 exam questions for English language course was used. The results of the study show that the machine learning methods together with linguistically-motivated features perform satisfactorily in the automatic classification based on cognitive levels of Bloom’s taxonomy.
|