FUZZY LOGIC FOR ASSESSING STUDENT LEARNING BEHAVIOUR AND ACADEMIC OUTCOMES: A SYSTEMATIC LITERATURE REVIEW

Abstract [English]

Purpose: Conventional, score-based assessment poorly captures the uncertainty, vagueness and subjectivity of student learning. Fuzzy logic, which models partial truth and degrees of membership, has been applied widely to evaluate and predict academic performance. This study systematically reviews that literature to consolidate methods, inputs and findings and to identify gaps relevant to assessing learning behaviour.
Design/Methodology/Approach: The systematic review was performed according to PRISMA 2020 guidelines. The studies in the peer-reviewed journals and conferences on the fuzzy logic evaluation of student performance were identified using major databases, screened with pre-defined eligibility criteria and synthesized narratively and thematically. Twenty-six studies between 2010 and 2025 were found that fulfilled the inclusion criteria and were plotted on context, method and fuzzy technique and findings.
Findings: Six themes were identified: the fuzzy inference pipeline (fuzzification, rule base, inference, defuzzification); input variables—in which academic scores were heavily relied on, and there was limited multidimensional behavioural measurement; fuzzy versus conventional grading—where fuzzy models were found to be consistently superior in terms of their interpretability and reduction in the number of misclassified students in the border region; hybrid and advanced approaches—such as neuro-fuzzy, clustering, fuzzy numbers and dynamic systems; application context—such as in distance learning and specific disciplines; and alignment with Outcome Based Education. The field methodologically privileges Mamdani-type inference and centroid defuzzification and seldom considers learning behaviour as a multidimensional construct in a structured manner or produces a composite behaviour score that is interpretable.
Originality/Value: The review brings together a disjointed landscape, outlines techniques and identifies gaps, and outlines an agenda for a fuzzy logic analytical framework that incorporates multidimensional learning behaviour into a composite Fuzzy Learning Behaviour Score that is relevant for a variety of higher-education contexts, including Qatar.