"Data for Optimized Instance Alteration for Explaining and Assessing Robustness of Classifiers"

"This repository contains the benchmark datasets used in the manuscript \u201cOptimized Instance Alteration for Explaining and Assessing Robustness of Classifiers\u201d The deposited data include publicly available image and tabular datasets, specifically MNIST, CIFAR-10, Coil2000, Phoneme, Wine, and the Breast Cancer dataset, which were used for empirical evaluation of the proposed method. These datasets support analysis of model misclassification, explanation through optimized input alteration, and robustness-related behavior across different data modalities. The repository is provided to document the data used in the study, support reproducibility of the reported experiments, and enable further research in explainable artificial intelligence, interpretability, and robustness assessment."