The robustness of explainable artificial intelligence (XAI) models is a cornerstone for their effectiveness and trustworthiness in high-stakes environments. As AI technologies permeate sensitive areas such as environmental monitoring and energy management, the demand for models that not only perform consistently but also offer transparent decision-making processes intensifies. This study introduces a transformative methodological innovation in XAI that significantly enhances the robustness and interpretability of predictive models. Employing advanced machine learning algorithms—XGBoost, Multi-layer Perceptron, and Keras Sequential Neural Networks—integrated with leading-edge XAI tools like Tree SHAP, Kernel SHAP, and Deep SHAP, our framework sets a new standard in the fidelity and accountability of AI systems.

Central to our methodology is the development of two novel metrics: the Stability Correlation Index (SCI) and the Explanation Integrity Metric (EIM). These metrics are specifically designed to critically evaluate the stability and reliability of explanations provided by AI models across varied and unpredictable data landscapes. By quantifying the consistency of model explanations under dynamic conditions, these metrics address a crucial gap in current AI practices—ensuring that AI systems remain robust against data perturbations and environmental changes, which is essential for their deployment in real-world scenarios.

Our application of this robust framework to a case study in Community Renewable Energy (CRE) illustrates its practical value. In CRE projects, understanding the multifaceted factors influencing community engagement is vital for the successful implementation and sustainability of renewable initiatives. By applying our robust XAI methods, we not only enhance the predictive accuracy but also provide stakeholders with clear, interpretable insights into the complex dynamics of community participation. This dual capability is indispensable for designing effective, data-driven strategies that promote widespread adoption and support for renewable energy projects.