The process of validating the accuracy of a classification by comparing it to ground truth data or other reference datasets.