DiCE-Extended: A Robust Approach to Counterfactual Explanations in Machine Learning