AUTOMATIC ANALYSIS OF EXAM PAPERS BASED ON BLOOM'S TAXONOMY: AN ARTIFICIAL INTELLIGENCE APPROACH

The quality assurance of the educational process cannot be achieved fully without evaluating all its elements, including the evaluation of the assessment tools itself.It is well known that exam papers are the predominant form of student assessment.To avoid using inadequate assessment of students' performance, many educational institutions resort to manually calibrating written exam papers based on specific qualitative models such as Bloom's model, which can be used to evaluate the different cognitive levels attained by learners.This model defines six cognitive levels that can be used to ensure a balanced assessment of student performance.This paper aims to develop an intelligent tool that can be used to automatically calibrate written exam papers based on the cognitive levels defined by Bloom taxonomy.This approach uses a combination of Bloom's taxonomy and natural language processing techniques.The supporting tool will contribute to ensuring the quality of written exam papers by subjecting learners to fair assessments and addressing any methodological errors in the process student evaluation.This research is supported by developing a tool that automatically classifies exam questions by extracting specific features using TF-IDF.A dataset of 600 questions was used to test the proposed approach.The data is divided equally among the different human cognition levels defined by Bloom.For classification purposes, the Support Vector Machine (SVM) classifier was used.The model was evaluated using the F1 score and its accuracy reached 80%, which is a very satisfactory result, and encourages further research.