Corn Yield Prediction in US Midwest Using Artificial Neural Networks

Climate change driven increment in temperature and variations in weather have affected agrarian economies throughout the world. Due to this temperature change-fueled uncertainty in agricultural yield, it becomes imperative to study the dependence of yield on meteorological factors. Deep learning architectures offer a way to clearly define this relationship through non-linear function approximations. In this study, we offer a comparison of deep learning with other popular data driven methods and outline a concrete dropout based Bayesian uncertainty estimation of yield predictions.