Utilizing machine learning models for accurate crop prediction and yield optimization

Precision agriculture (PA) represents one of the most promising applications of Artificial Intelligence (AI) and Internet of Things (IoT) technologies, and its practical implementations often rely on data collected from sensors such as weather stations, cameras, or leaf- and soil-sampling devices. Accurate prediction of crop yield is essential for agricultural producers, companies, and nations that are interested in food supply chain and agricultural planning, not just to optimize profit margins for producers but also to ensure adequate food security for the population. During the past three decades, machine learning models have become the primary way to address crop yield prediction tasks, owing to their efficient leveraging of large historical crop yield datasets in modeling complex nonlinear functions between exogenous factors and yield. In this paper, we examine this large body of literature on traditional and recently proposed machine learning techniques and how they succeed in predicting crop yield evaluated on different crops and countries.