Published Paper Details:

Modern agriculture faces critical challenges in meeting rising food demand amid resource constraints and climate change. Advanced machine learning (ML) techniques--particularly algorithms like Random Forest (RF), Support Vector Machine (SVM), and Decision Trees--can capture complex, high-dimensional interactions among climate, soil, and management factors. Moreover, the integration of IoT devices and remote sensing provides real-time data that further enhances prediction accuracy. In this survey, we systematically review recent studies (2017-2025) on ML-driven crop yield prediction and resource optimization. The evidence shows that ML models frequently achieve high accuracy (often >90%) in yield or soil moisture prediction. Key insights include the importance of environmental features and advanced algorithms. Despite these successes, gaps remain in real-time field validation, data diversity, and standard benchmarking. This review highlights how ML-based approaches can significantly improve crop yield forecasting and resource use efficiency, providing valuable decision support for sustainable agriculture.