MindGauge: A GUI-Based Machine Learning Tool for EEG Workload Classification

Le résumé fourni par la source

This paper MindGauge, a graphical user interface (GUI) that enables non-technical users to assess mental workload from EEG with a simple upload-and-analyze workflow. Users provide. mat or. csv EEG files; the system performs automated preprocessing, band-power feature extraction (delta, theta, alpha, beta, gamma), and classification using Support Vector Machines (SVMs). Two models are available: binary (low vs. high workload; linear SVM) and multiclass (low/medium/high; polynomial SVM), both evaluated with Leave-One-Group-Out (subject-wise) cross-validation for subject-independent testing. We trained on 50 subjects performing N-back tasks (0-, 1-, 2-back) and validated end-toend through the GUI. The GUI also supports pretrained model upload, session history, and PDF report export for practical use. MindGauge demonstrates that a lightweight, interpretable pipeline can deliver interactive workload estimates while remaining accessible to non-experts. We discuss design choices, dataset and features, and provide a compact comparison to recent EEG workload studies.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.