Aulia Yolanda Ritonga
Universitas Negeri Medan

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

PENERAPAN LOGIKA FUZZY DALAM PEMBELAJARAN DAN EVALUASI PENDIDIKAN: SYSTEMATIC LITERATURE REVIEW Aulia Yolanda Ritonga; Mutiara Arini Putri; Grace Patresya Tambunan; Putra Rodoando Marudut Pardede; Nuelsen Leo Metro Sihombing; Irwan Sanjaya; Elsa Sabrina
Didaktik : Jurnal Ilmiah PGSD STKIP Subang Vol. 12 No. 04 (2026): Volume 12 No. 4, Desember 2026 Released
Publisher : STKIP Subang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36989/didaktik.v12i04.19876

Abstract

Fuzzy logic has been increasingly used in educational systems to represent gradations, ambiguity, and multiple indicators that are difficult to capture through rigid threshold-based assessment. This review synthesizes 20 unique full-text articles supplied for analysis concerning fuzzy logic in learning and educational evaluation. The synthesis addresses research characteristics, learning applications, assessment applications, fuzzy methods, reported benefits and limitations, and research gaps. The reviewed studies cover adaptive e-learning and intelligent tutoring, readiness and performance prediction, competency assessment, online course quality evaluation, and multi-criteria decision support. Mamdani and Sugeno fuzzy inference systems are prominent in score estimation and performance assessment, while fuzzy Delphi, fuzzy DEMATEL, AHP–fuzzy comprehensive evaluation, and fuzzy clustering are used for validation, factor analysis, and personalization. Reported benefits include interpretable rule-based decisions, integration of heterogeneous indicators, and more nuanced treatment of borderline cases. However, small or institution-specific samples, dependence on expert-defined membership functions and rules, limited external validation, and unclear separation between model effects and instructional interventions constrain the evidence. Fuzzy logic should therefore be understood as a decision-support approach rather than an automatic guarantee of accuracy, fairness, or improved learning. Future research should prioritize multi-site validation, transparent reporting, comparisons with baseline models, fairness analysis, and longitudinal studies connecting model outputs to learning processes and outcomes.