Classification of Bloom’s Taxonomy Based University Examination Questions: A Recurrent Neural Network and Natural Language Processing Approach

The true goal of learning and teaching can be effectively achieved through written examinations, which serve as a standard method to evaluate learners’ knowledge and understanding. Hence, it is essential to meticulously select well-crafted questions that evaluate diverse cognitive abilities at different levels. Creating appropriately labeled and well-structured question papers is a challenging task. Presently the low employability of recent engineering graduates and the absence of top-tier engineering institutions in global rankings highlight the issue of poor-quality engineering education in India, emphasizing the need for defining “quality” in terms of learning outcomes. Bloom's Taxonomy (BT) has gained widespread acceptance for categorizing questions, offering valuable guidance in crafting examination questions across various engineering institutions. This paper suggests an automated examination question analysis to categorize questions based on this taxonomy. However, the results of our proposed work using Recurrent Neural Network (RNN) seem to be quite satisfactory.