Explainable Brain Tumor Detection Using EfficientNetV2-S and RBF-SVM: A Multi-Source MRI Study — Reproducibility Archive

This reproducibility archive accompanies the study: "Explainable Brain Tumor Detection Using EfficientNetV2-S and RBF-SVM: A Multi-Source MRI Study." The archive provides the computational resources required to reproduce the reported analyses, including dataset provenance documentation, integrity auditing records, frozen evaluation protocols, model information, prediction outputs, statistical analyses, and explainable artificial intelligence (XAI) results. The proposed framework uses frozen EfficientNetV2-S feature extraction, PCA-based dimensionality reduction, and calibrated RBF-SVM classification with source-held-out evaluation. Original MRI images are not redistributed because they remain subject to the licensing restrictions of their respective source datasets. This archive provides metadata, manifests, audit records, models, outputs, and analysis materials required for computational reproducibility. This release represents computational validation and does not establish clinical deployment readiness or prospective clinical validation.