EzJSU: An Intelligent Decision Support System for Outcome-Based Assessment Auditing and Bloom’s Taxonomy Realignment

About the Project

Manual cross-auditing of examination papers against Test Specification Blueprints (Jadual Spesifikasi Ujian / JSU) in Outcome-Based Education (OBE) is labor-intensive and vulnerable to subjective classification errors. This study presents EZJSU, a web-based decision support system leveraging generative language models and optical character recognition to automate assessment auditing against Malaysian Higher Education standards (MQA/UiTM). The platform ingests multi-modal assessment inputs, classifies cognitive complexity (C1–C6), evaluates Course Learning Outcome (CLO) mappings, and identifies alignment discrepancies. Evaluated across objective and subjective paper formats, AuditAI reduces auditing throughput times by over 90% while providing dynamic, real-time question realignment and rubrics generation.