Faiz Hilmawan Masyfa, Daniel Siahaan, Umi Laili Yuhana
High-stakes game-based assessment needs adaptation that supports students with different ability levels while maintaining score interpretation. Current systems often adjust assessment parameters without a clear link between learner profiles and control policy. This paper proposes an auditable decision support system that uses a directional rulebook to map cognitive and motivational profiles into adaptive assessment parameter changes. Tabular profiling models act as sensors, and SHAP coverage links explainable model to rules. In a crossover deployment with 102 paired observations, the expert-reviewed rulebook maintained comparable scores and completion time while significantly reducing the number of completed items before the endpoint. A comparison of seven profiling models using stratified 10-fold cross-validation showed that SE-TabNet achieved the best macro-F1 for Ep, Cf, Te, and Me with 0.804, 0.789, 0.666, and 0.951, while TabNet performed best for Ps and Ac with 0.976 and 0.913. These findings support auditable rule-based adaptation in high-stakes game-based assessment. This article is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. License details: https://creativecommons.org/licenses/by-sa/4.0/
Department of Informatics, Institut Teknologi Sepuluh Nopember, Surabaya, Indonesia; Department of Electrical Engineering and Informatics, Universitas Negeri Malang, Malang, Indonesia
Research at a Glance
Register to unlock
Topics & SDG Alignment
Register to unlock
Collaboration
Register to unlock
Author Profile (Selected)
Register to unlock
References Overview
Register to unlock
Journal & Source
Register to unlock
Metadata & Integrity
Register to unlock