Machine Learning Approaches for Climate-Smart Crop Recommendation

Agriculture is highly vulnerable to climate variability, with studies indicating that climate change could reduce global crop yields by up to 10–25% by 2050, particularly in developing countries. This has intensified the need for intelligent and adaptive crop recommendation systems that can respond to dynamic environmental conditions. Traditional approaches primarily rely on static soil parameters and fail to incorporate real-time weather factors such as temperature, humidity, and rainfall, resulting in suboptimal decision-making. This paper presents a comprehensive review of machine learning approaches for climate-smart crop recommendation systems. It analyzes recent advancements in artificial intelligence techniques, including Random Forest, Support Vector Machine (SVM), and XGBoost, highlighting their effectiveness in handling complex agricultural datasets. The review also examines the role of feature engineering, particularly the development of composite indicators such as the Climate Risk Index (CRI), which integrates multiple environmental variables to enhance prediction accuracy and robustness. Furthermore, this study evaluates the integration of multi-source data, including soil characteristics and meteorological information, to improve system performance under varying climatic conditions. Comparative analysis of existing models indicates that hybrid and ensemble approaches consistently outperform traditional methods, achieving accuracy levels exceeding 90–95% in crop prediction tasks. The review identifies key challenges such as a lack of real-time data integration, limited climate resilience, and poor model interpretability. It also outlines future research directions, including the incorporation of explainable AI, IoT-based real-time monitoring, and scalable deployment frameworks. Overall, this paper provides a comprehensive understanding of climate-smart crop recommendation systems and contributes to the development of sustainable, data-driven agricultural practices.