Explainable AI and Hybrid Machine Learning Approaches for Crop Yield Prediction: A Systematic Review

Accurate crop yield assessment is essential for climate-resilient farming methods, effective agricultural resource allocation, and food security.By discovering intricate nonlinear correlations between climatic, soil, and remote sensing factors, recent advances in machine learning (ML) and deep learning (DL) have significantly increased the accuracy of crop yield estimation. Accuracy and convergence robustness are further enhanced by hybrid machine learning models that include ensemble techniques and optimization algorithms.The majority of correct models, however, are black-box models, which reduces their interpretability, reliability, and utility.Explainable Artificial Intelligence (XAI) and hybrid machine learning models for crop yield estimation are systematically reviewed in this research.A PRISMA-based framework was used to survey and categorize papers published between 2017 and 2025 into four groups: hybrid optimization-based models, XAIbased models, deep learning models, and classical machine learning models.The findings indicate that although hybrid and ensemble models are the most accurate, the use of interpretability techniques like SHAP and LIME is still relatively new