Enhancing Solar Power Forecasting in Abuja: An Application of XAI Techniques for Improved Transparency and Trust

The increasing integration of photovoltaic (PV) power into energy grids necessitates accurate forecasting to enhance grid stability and optimize energy management. However, traditional machine learning models often lack interpretability, leading to concerns regarding reliability and trust in AI-driven predictions. This study applies Explainable AI (XAI) techniques, including SHapley Additive exPlanations (SHAP), Local Interpretable Model-agnostic Explanations (LIME), and feature importance plots, to improve the transparency of a hybrid Least Absolute Shrinkage and Selection Operator-Random Forest Regression (LASSO-RFR) model for PV power forecasting in Abuja, Nigeria. By analyzing meteorological datasets from 2020–2022, the hybrid model was evaluated using key performance metrics such as Root Mean Squared Error (RMSE), and R-squared <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(\mathbf{R}^{2})$</tex>. The results indicate that the hybrid LASSO-RFR framework enhances prediction accuracy while maintaining model interpretability. XAI analyses revealed that surface pressure, dew point temperature, and air temperature were the most influential variables in forecasting solar radiation. The findings underscore the importance of integrating XAI tools into AI-driven energy forecasting models to improve transparency, stakeholder confidence, and practical deployment in real-world scenarios.